----------------------\n HAProxy\n Configuration Manual\n ----------------------\n version 1.4.22\n willy tarreau\n 2012/08/14\n\n\nThis document covers the configuration language as implemented in the version\nspecified above. It does not provide any hint, example or advice. For such\ndocumentation, please refer to the Reference Manual or the Architecture Manual.\nThe summary below is meant to help you search sections by name and navigate\nthrough the document.\n\nNote to documentation contributors :\n This document is formated with 80 columns per line, with even number of\n spaces for indentation and without tabs. Please follow these rules strictly\n so that it remains easily printable everywhere. If a line needs to be\n printed verbatim and does not fit, please end each line with a backslash\n ('\') and continue on next line. If you add sections, please update the\n summary below for easier searching.\n\n\nSummary\n-------\n\n1. Quick reminder about HTTP\n1.1. The HTTP transaction model\n1.2. HTTP request\n1.2.1. The Request line\n1.2.2. The request headers\n1.3. HTTP response\n1.3.1. The Response line\n1.3.2. The response headers\n\n2. Configuring HAProxy\n2.1. Configuration file format\n2.2. Time format\n2.3. Examples\n\n3. Global parameters\n3.1. Process management and security\n3.2. Performance tuning\n3.3. Debugging\n3.4. Userlists\n\n4. Proxies\n4.1. Proxy keywords matrix\n4.2. Alphabetically sorted keywords reference\n\n5. Server and default-server options\n\n6. HTTP header manipulation\n\n7. Using ACLs and pattern extraction\n7.1. Matching integers\n7.2. Matching strings\n7.3. Matching regular expressions (regexes)\n7.4. Matching IPv4 addresses\n7.5. Available matching criteria\n7.5.1. Matching at Layer 4 and below\n7.5.2. Matching contents at Layer 4\n7.5.3. Matching at Layer 7\n7.6. Pre-defined ACLs\n7.7. Using ACLs to form conditions\n7.8. Pattern extraction\n\n8. Logging\n8.1. Log levels\n8.2. Log formats\n8.2.1. Default log format\n8.2.2. TCP log format\n8.2.3. HTTP log format\n8.3. Advanced logging options\n8.3.1. Disabling logging of external tests\n8.3.2. Logging before waiting for the session to terminate\n8.3.3. Raising log level upon errors\n8.3.4. Disabling logging of successful connections\n8.4. Timing events\n8.5. Session state at disconnection\n8.6. Non-printable characters\n8.7. Capturing HTTP cookies\n8.8. Capturing HTTP headers\n8.9. Examples of logs\n\n9. Statistics and monitoring\n9.1. CSV format\n9.2. Unix Socket commands\n\n\n1. Quick reminder about HTTP\n----------------------------\n\nWhen haproxy is running in HTTP mode, both the request and the response are\nfully analyzed and indexed, thus it becomes possible to build matching criteria\non almost anything found in the contents.\n\nHowever, it is important to understand how HTTP requests and responses are\nformed, and how HAProxy decomposes them. It will then become easier to write\ncorrect rules and to debug existing configurations.\n\n\n1.1. The HTTP transaction model\n-------------------------------\n\nThe HTTP protocol is transaction-driven. This means that each request will lead\nto one and only one response. Traditionally, a TCP connection is established\nfrom the client to the server, a request is sent by the client on the\nconnection, the server responds and the connection is closed. A new request\nwill involve a new connection :\n\n [CON1] [REQ1] ... [RESP1] [CLO1] [CON2] [REQ2] ... [RESP2] [CLO2] ...\n\nIn this mode, called the "HTTP close" mode, there are as many connection\nestablishments as there are HTTP transactions. Since the connection is closed\nby the server after the response, the client does not need to know the content\nlength.\n\nDue to the transactional nature of the protocol, it was possible to improve it\nto avoid closing a connection between two subsequent transactions. In this mode\nhowever, it is mandatory that the server indicates the content length for each\nresponse so that the client does not wait indefinitely. For this, a special\nheader is used: "Content-length". This mode is called the "keep-alive" mode :\n\n [CON] [REQ1] ... [RESP1] [REQ2] ... [RESP2] [CLO] ...\n\nIts advantages are a reduced latency between transactions, and less processing\npower required on the server side. It is generally better than the close mode,\nbut not always because the clients often limit their concurrent connections to\na smaller value.\n\nA last improvement in the communications is the pipelining mode. It still uses\nkeep-alive, but the client does not wait for the first response to send the\nsecond request. This is useful for fetching large number of images composing a\npage :\n\n [CON] [REQ1] [REQ2] ... [RESP1] [RESP2] [CLO] ...\n\nThis can obviously have a tremendous benefit on performance because the network\nlatency is eliminated between subsequent requests. Many HTTP agents do not\ncorrectly support pipelining since there is no way to associate a response with\nthe corresponding request in HTTP. For this reason, it is mandatory for the\nserver to reply in the exact same order as the requests were received.\n\nBy default HAProxy operates in a tunnel-like mode with regards to persistent\nconnections: for each connection it processes the first request and forwards\neverything else (including additional requests) to selected server. Once\nestablished, the connection is persisted both on the client and server\nsides. Use "option http-server-close" to preserve client persistent connections\nwhile handling every incoming request individually, dispatching them one after\nanother to servers, in HTTP close mode. Use "option httpclose" to switch both\nsides to HTTP close mode. "option forceclose" and "option\nhttp-pretend-keepalive" help working around servers misbehaving in HTTP close\nmode.\n\n\n1.2. HTTP request\n-----------------\n\nFirst, let's consider this HTTP request :\n\n Line Contents\n number\n 1 GET /serv/login.php?lang=en&profile=2 HTTP/1.1\n 2 Host: www.mydomain.com\n 3 User-agent: my small browser\n 4 Accept: image/jpeg, image/gif\n 5 Accept: image/png\n\n\n1.2.1. The Request line\n-----------------------\n\nLine 1 is the "request line". It is always composed of 3 fields :\n\n - a METHOD : GET\n - a URI : /serv/login.php?lang=en&profile=2\n - a version tag : HTTP/1.1\n\nAll of them are delimited by what the standard calls LWS (linear white spaces),\nwhich are commonly spaces, but can also be tabs or line feeds/carriage returns\nfollowed by spaces/tabs. The method itself cannot contain any colon (':') and\nis limited to alphabetic letters. All those various combinations make it\ndesirable that HAProxy performs the splitting itself rather than leaving it to\nthe user to write a complex or inaccurate regular expression.\n\nThe URI itself can have several forms :\n\n - A "relative URI" :\n\n /serv/login.php?lang=en&profile=2\n\n It is a complete URL without the host part. This is generally what is\n received by servers, reverse proxies and transparent proxies.\n\n - An "absolute URI", also called a "URL" :\n\n http://192.168.0.12:8080/serv/login.php?lang=en&profile=2\n\n It is composed of a "scheme" (the protocol name followed by '://'), a host\n name or address, optionally a colon (':') followed by a port number, then\n a relative URI beginning at the first slash ('/') after the address part.\n This is generally what proxies receive, but a server supporting HTTP/1.1\n must accept this form too.\n\n - a star ('*') : this form is only accepted in association with the OPTIONS\n method and is not relayable. It is used to inquiry a next hop's\n capabilities.\n\n - an address:port combination : 192.168.0.12:80\n This is used with the CONNECT method, which is used to establish TCP\n tunnels through HTTP proxies, generally for HTTPS, but sometimes for\n other protocols too.\n\nIn a relative URI, two sub-parts are identified. The part before the question\nmark is called the "path". It is typically the relative path to static objects\non the server. The part after the question mark is called the "query string".\nIt is mostly used with GET requests sent to dynamic scripts and is very\nspecific to the language, framework or application in use.\n\n\n1.2.2. The request headers\n--------------------------\n\nThe headers start at the second line. They are composed of a name at the\nbeginning of the line, immediately followed by a colon (':'). Traditionally,\nan LWS is added after the colon but that's not required. Then come the values.\nMultiple identical headers may be folded into one single line, delimiting the\nvalues with commas, provided that their order is respected. This is commonly\nencountered in the "Cookie:" field. A header may span over multiple lines if\nthe subsequent lines begin with an LWS. In the example in 1.2, lines 4 and 5\ndefine a total of 3 values for the "Accept:" header.\n\nContrary to a common mis-conception, header names are not case-sensitive, and\ntheir values are not either if they refer to other header names (such as the\n"Connection:" header).\n\nThe end of the headers is indicated by the first empty line. People often say\nthat it's a double line feed, which is not exact, even if a double line feed\nis one valid form of empty line.\n\nFortunately, HAProxy takes care of all these complex combinations when indexing\nheaders, checking values and counting them, so there is no reason to worry\nabout the way they could be written, but it is important not to accuse an\napplication of being buggy if it does unusual, valid things.\n\nImportant note:\n As suggested by RFC2616, HAProxy normalizes headers by replacing line breaks\n in the middle of headers by LWS in order to join multi-line headers. This\n is necessary for proper analysis and helps less capable HTTP parsers to work\n correctly and not to be fooled by such complex constructs.\n\n\n1.3. HTTP response\n------------------\n\nAn HTTP response looks very much like an HTTP request. Both are called HTTP\nmessages. Let's consider this HTTP response :\n\n Line Contents\n number\n 1 HTTP/1.1 200 OK\n 2 Content-length: 350\n 3 Content-Type: text/html\n\nAs a special case, HTTP supports so called "Informational responses" as status\ncodes 1xx. These messages are special in that they don't convey any part of the\nresponse, they're just used as sort of a signaling message to ask a client to\ncontinue to post its request for instance. In the case of a status 100 response\nthe requested information will be carried by the next non-100 response message\nfollowing the informational one. This implies that multiple responses may be\nsent to a single request, and that this only works when keep-alive is enabled\n(1xx messages are HTTP/1.1 only). HAProxy handles these messages and is able to\ncorrectly forward and skip them, and only process the next non-100 response. As\nsuch, these messages are neither logged nor transformed, unless explicitly\nstate otherwise. Status 101 messages indicate that the protocol is changing\nover the same connection and that haproxy must switch to tunnel mode, just as\nif a CONNECT had occurred. Then the Upgrade header would contain additional\ninformation about the type of protocol the connection is switching to.\n\n\n1.3.1. The Response line\n------------------------\n\nLine 1 is the "response line". It is always composed of 3 fields :\n\n - a version tag : HTTP/1.1\n - a status code : 200\n - a reason : OK\n\nThe status code is always 3-digit. The first digit indicates a general status :\n - 1xx = informational message to be skipped (eg: 100, 101)\n - 2xx = OK, content is following (eg: 200, 206)\n - 3xx = OK, no content following (eg: 302, 304)\n - 4xx = error caused by the client (eg: 401, 403, 404)\n - 5xx = error caused by the server (eg: 500, 502, 503)\n\nPlease refer to RFC2616 for the detailed meaning of all such codes. The\n"reason" field is just a hint, but is not parsed by clients. Anything can be\nfound there, but it's a common practice to respect the well-established\nmessages. It can be composed of one or multiple words, such as "OK", "Found",\nor "Authentication Required".\n\nHaproxy may emit the following status codes by itself :\n\n Code When / reason\n 200 access to stats page, and when replying to monitoring requests\n 301 when performing a redirection, depending on the configured code\n 302 when performing a redirection, depending on the configured code\n 303 when performing a redirection, depending on the configured code\n 400 for an invalid or too large request\n 401 when an authentication is required to perform the action (when\n accessing the stats page)\n 403 when a request is forbidden by a "block" ACL or "reqdeny" filter\n 408 when the request timeout strikes before the request is complete\n 500 when haproxy encounters an unrecoverable internal error, such as a\n memory allocation failure, which should never happen\n 502 when the server returns an empty, invalid or incomplete response, or\n when an "rspdeny" filter blocks the response.\n 503 when no server was available to handle the request, or in response to\n monitoring requests which match the "monitor fail" condition\n 504 when the response timeout strikes before the server responds\n\nThe error 4xx and 5xx codes above may be customized (see "errorloc" in section\n4.2).\n\n\n1.3.2. The response headers\n---------------------------\n\nResponse headers work exactly like request headers, and as such, HAProxy uses\nthe same parsing function for both. Please refer to paragraph 1.2.2 for more\ndetails.\n\n\n2. Configuring HAProxy\n----------------------\n\n2.1. Configuration file format\n------------------------------\n\nHAProxy's configuration process involves 3 major sources of parameters :\n\n - the arguments from the command-line, which always take precedence\n - the "global" section, which sets process-wide parameters\n - the proxies sections which can take form of "defaults", "listen",\n "frontend" and "backend".\n\nThe configuration file syntax consists in lines beginning with a keyword\nreferenced in this manual, optionally followed by one or several parameters\ndelimited by spaces. If spaces have to be entered in strings, then they must be\npreceded by a backslash ('\') to be escaped. Backslashes also have to be\nescaped by doubling them.\n\n\n2.2. Time format\n----------------\n\nSome parameters involve values representing time, such as timeouts. These\nvalues are generally expressed in milliseconds (unless explicitly stated\notherwise) but may be expressed in any other unit by suffixing the unit to the\nnumeric value. It is important to consider this because it will not be repeated\nfor every keyword. Supported units are :\n\n - us : microseconds. 1 microsecond = 1/1000000 second\n - ms : milliseconds. 1 millisecond = 1/1000 second. This is the default.\n - s : seconds. 1s = 1000ms\n - m : minutes. 1m = 60s = 60000ms\n - h : hours. 1h = 60m = 3600s = 3600000ms\n - d : days. 1d = 24h = 1440m = 86400s = 86400000ms\n\n\n2.3. Examples\n-------------\n\n # Simple configuration for an HTTP proxy listening on port 80 on all\n # interfaces and forwarding requests to a single backend "servers" with a\n # single server "server1" listening on 127.0.0.1:8000\n global\n daemon\n maxconn 256\n\n defaults\n mode http\n timeout connect 5000ms\n timeout client 50000ms\n timeout server 50000ms\n\n frontend http-in\n bind *:80\n default_backend servers\n\n backend servers\n server server1 127.0.0.1:8000 maxconn 32\n\n\n # The same configuration defined with a single listen block. Shorter but\n # less expressive, especially in HTTP mode.\n global\n daemon\n maxconn 256\n\n defaults\n mode http\n timeout connect 5000ms\n timeout client 50000ms\n timeout server 50000ms\n\n listen http-in\n bind *:80\n server server1 127.0.0.1:8000 maxconn 32\n\n\nAssuming haproxy is in $PATH, test these configurations in a shell with:\n\n $ sudo haproxy -f configuration.conf -c\n\n\n3. Global parameters\n--------------------\n\nParameters in the "global" section are process-wide and often OS-specific. They\nare generally set once for all and do not need being changed once correct. Some\nof them have command-line equivalents.\n\nThe following keywords are supported in the "global" section :\n\n * Process management and security\n - chroot\n - daemon\n - gid\n - group\n - log\n - log-send-hostname\n - nbproc\n - pidfile\n - uid\n - ulimit-n\n - user\n - stats\n - node\n - description\n\n * Performance tuning\n - maxconn\n - maxpipes\n - noepoll\n - nokqueue\n - nopoll\n - nosepoll\n - nosplice\n - spread-checks\n - tune.bufsize\n - tune.chksize\n - tune.maxaccept\n - tune.maxpollevents\n - tune.maxrewrite\n - tune.rcvbuf.client\n - tune.rcvbuf.server\n - tune.sndbuf.client\n - tune.sndbuf.server\n\n * Debugging\n - debug\n - quiet\n\n\n3.1. Process management and security\n------------------------------------\n\nchroot <jail dir>\n Changes current directory to <jail dir> and performs a chroot() there before\n dropping privileges. This increases the security level in case an unknown\n vulnerability would be exploited, since it would make it very hard for the\n attacker to exploit the system. This only works when the process is started\n with superuser privileges. It is important to ensure that <jail_dir> is both\n empty and unwritable to anyone.\n\ndaemon\n Makes the process fork into background. This is the recommended mode of\n operation. It is equivalent to the command line "-D" argument. It can be\n disabled by the command line "-db" argument.\n\ngid <number>\n Changes the process' group ID to <number>. It is recommended that the group\n ID is dedicated to HAProxy or to a small set of similar daemons. HAProxy must\n be started with a user belonging to this group, or with superuser privileges.\n See also "group" and "uid".\n\ngroup <group name>\n Similar to "gid" but uses the GID of group name <group name> from /etc/group.\n See also "gid" and "user".\n\nlog <address> <facility> [max level [min level]]\n Adds a global syslog server. Up to two global servers can be defined. They\n will receive logs for startups and exits, as well as all logs from proxies\n configured with "log global".\n\n <address> can be one of:\n\n - An IPv4 address optionally followed by a colon and a UDP port. If\n no port is specified, 514 is used by default (the standard syslog\n port).\n\n - A filesystem path to a UNIX domain socket, keeping in mind\n considerations for chroot (be sure the path is accessible inside\n the chroot) and uid/gid (be sure the path is appropriately\n writeable).\n\n <facility> must be one of the 24 standard syslog facilities :\n\n kern user mail daemon auth syslog lpr news\n uucp cron auth2 ftp ntp audit alert cron2\n local0 local1 local2 local3 local4 local5 local6 local7\n\n An optional level can be specified to filter outgoing messages. By default,\n all messages are sent. If a maximum level is specified, only messages with a\n severity at least as important as this level will be sent. An optional minimum\n level can be specified. If it is set, logs emitted with a more severe level\n than this one will be capped to this level. This is used to avoid sending\n "emerg" messages on all terminals on some default syslog configurations.\n Eight levels are known :\n\n emerg alert crit err warning notice info debug\n\nlog-send-hostname [<string>]\n Sets the hostname field in the syslog header. If optional "string" parameter\n is set the header is set to the string contents, otherwise uses the hostname\n of the system. Generally used if one is not relaying logs through an\n intermediate syslog server or for simply customizing the hostname printed in\n the logs.\n\nlog-tag <string>\n Sets the tag field in the syslog header to this string. It defaults to the\n program name as launched from the command line, which usually is "haproxy".\n Sometimes it can be useful to differentiate between multiple processes\n running on the same host.\n\nnbproc <number>\n Creates <number> processes when going daemon. This requires the "daemon"\n mode. By default, only one process is created, which is the recommended mode\n of operation. For systems limited to small sets of file descriptors per\n process, it may be needed to fork multiple daemons. USING MULTIPLE PROCESSES\n IS HARDER TO DEBUG AND IS REALLY DISCOURAGED. See also "daemon".\n\npidfile <pidfile>\n Writes pids of all daemons into file <pidfile>. This option is equivalent to\n the "-p" command line argument. The file must be accessible to the user\n starting the process. See also "daemon".\n\nstats socket <path> [{uid | user} <uid>] [{gid | group} <gid>] [mode <mode>]\n [level <level>]\n\n Creates a UNIX socket in stream mode at location <path>. Any previously\n existing socket will be backed up then replaced. Connections to this socket\n will return various statistics outputs and even allow some commands to be\n issued. Please consult section 9.2 "Unix Socket commands" for more details.\n\n An optional "level" parameter can be specified to restrict the nature of\n the commands that can be issued on the socket :\n - "user" is the least privileged level ; only non-sensitive stats can be\n read, and no change is allowed. It would make sense on systems where it\n is not easy to restrict access to the socket.\n\n - "operator" is the default level and fits most common uses. All data can\n be read, and only non-sensitive changes are permitted (eg: clear max\n counters).\n\n - "admin" should be used with care, as everything is permitted (eg: clear\n all counters).\n\n On platforms which support it, it is possible to restrict access to this\n socket by specifying numerical IDs after "uid" and "gid", or valid user and\n group names after the "user" and "group" keywords. It is also possible to\n restrict permissions on the socket by passing an octal value after the "mode"\n keyword (same syntax as chmod). Depending on the platform, the permissions on\n the socket will be inherited from the directory which hosts it, or from the\n user the process is started with.\n\nstats timeout <timeout, in milliseconds>\n The default timeout on the stats socket is set to 10 seconds. It is possible\n to change this value with "stats timeout". The value must be passed in\n milliseconds, or be suffixed by a time unit among { us, ms, s, m, h, d }.\n\nstats maxconn <connections>\n By default, the stats socket is limited to 10 concurrent connections. It is\n possible to change this value with "stats maxconn".\n\nuid <number>\n Changes the process' user ID to <number>. It is recommended that the user ID\n is dedicated to HAProxy or to a small set of similar daemons. HAProxy must\n be started with superuser privileges in order to be able to switch to another\n one. See also "gid" and "user".\n\nulimit-n <number>\n Sets the maximum number of per-process file-descriptors to <number>. By\n default, it is automatically computed, so it is recommended not to use this\n option.\n\nuser <user name>\n Similar to "uid" but uses the UID of user name <user name> from /etc/passwd.\n See also "uid" and "group".\n\nnode <name>\n Only letters, digits, hyphen and underscore are allowed, like in DNS names.\n\n This statement is useful in HA configurations where two or more processes or\n servers share the same IP address. By setting a different node-name on all\n nodes, it becomes easy to immediately spot what server is handling the\n traffic.\n\ndescription <text>\n Add a text that describes the instance.\n\n Please note that it is required to escape certain characters (# for example)\n and this text is inserted into a html page so you should avoid using\n "<" and ">" characters.\n\n\n3.2. Performance tuning\n-----------------------\n\nmaxconn <number>\n Sets the maximum per-process number of concurrent connections to <number>. It\n is equivalent to the command-line argument "-n". Proxies will stop accepting\n connections when this limit is reached. The "ulimit-n" parameter is\n automatically adjusted according to this value. See also "ulimit-n".\n\nmaxpipes <number>\n Sets the maximum per-process number of pipes to <number>. Currently, pipes\n are only used by kernel-based tcp splicing. Since a pipe contains two file\n descriptors, the "ulimit-n" value will be increased accordingly. The default\n value is maxconn/4, which seems to be more than enough for most heavy usages.\n The splice code dynamically allocates and releases pipes, and can fall back\n to standard copy, so setting this value too low may only impact performance.\n\nnoepoll\n Disables the use of the "epoll" event polling system on Linux. It is\n equivalent to the command-line argument "-de". The next polling system\n used will generally be "poll". See also "nosepoll", and "nopoll".\n\nnokqueue\n Disables the use of the "kqueue" event polling system on BSD. It is\n equivalent to the command-line argument "-dk". The next polling system\n used will generally be "poll". See also "nopoll".\n\nnopoll\n Disables the use of the "poll" event polling system. It is equivalent to the\n command-line argument "-dp". The next polling system used will be "select".\n It should never be needed to disable "poll" since it's available on all\n platforms supported by HAProxy. See also "nosepoll", and "nopoll" and\n "nokqueue".\n\nnosepoll\n Disables the use of the "speculative epoll" event polling system on Linux. It\n is equivalent to the command-line argument "-ds". The next polling system\n used will generally be "epoll". See also "nosepoll", and "nopoll".\n\nnosplice\n Disables the use of kernel tcp splicing between sockets on Linux. It is\n equivalent to the command line argument "-dS". Data will then be copied\n using conventional and more portable recv/send calls. Kernel tcp splicing is\n limited to some very recent instances of kernel 2.6. Most versions between\n 2.6.25 and 2.6.28 are buggy and will forward corrupted data, so they must not\n be used. This option makes it easier to globally disable kernel splicing in\n case of doubt. See also "option splice-auto", "option splice-request" and\n "option splice-response".\n\nspread-checks <0..50, in percent>\n Sometimes it is desirable to avoid sending health checks to servers at exact\n intervals, for instance when many logical servers are located on the same\n physical server. With the help of this parameter, it becomes possible to add\n some randomness in the check interval between 0 and +/- 50%. A value between\n 2 and 5 seems to show good results. The default value remains at 0.\n\ntune.bufsize <number>\n Sets the buffer size to this size (in bytes). Lower values allow more\n sessions to coexist in the same amount of RAM, and higher values allow some\n applications with very large cookies to work. The default value is 16384 and\n can be changed at build time. It is strongly recommended not to change this\n from the default value, as very low values will break some services such as\n statistics, and values larger than default size will increase memory usage,\n possibly causing the system to run out of memory. At least the global maxconn\n parameter should be decreased by the same factor as this one is increased.\n\ntune.chksize <number>\n Sets the check buffer size to this size (in bytes). Higher values may help\n find string or regex patterns in very large pages, though doing so may imply\n more memory and CPU usage. The default value is 16384 and can be changed at\n build time. It is not recommended to change this value, but to use better\n checks whenever possible.\n\ntune.maxaccept <number>\n Sets the maximum number of consecutive accepts that a process may perform on\n a single wake up. High values give higher priority to high connection rates,\n while lower values give higher priority to already established connections.\n This value is limited to 100 by default in single process mode. However, in\n multi-process mode (nbproc > 1), it defaults to 8 so that when one process\n wakes up, it does not take all incoming connections for itself and leaves a\n part of them to other processes. Setting this value to -1 completely disables\n the limitation. It should normally not be needed to tweak this value.\n\ntune.maxpollevents <number>\n Sets the maximum amount of events that can be processed at once in a call to\n the polling system. The default value is adapted to the operating system. It\n has been noticed that reducing it below 200 tends to slightly decrease\n latency at the expense of network bandwidth, and increasing it above 200\n tends to trade latency for slightly increased bandwidth.\n\ntune.maxrewrite <number>\n Sets the reserved buffer space to this size in bytes. The reserved space is\n used for header rewriting or appending. The first reads on sockets will never\n fill more than bufsize-maxrewrite. Historically it has defaulted to half of\n bufsize, though that does not make much sense since there are rarely large\n numbers of headers to add. Setting it too high prevents processing of large\n requests or responses. Setting it too low prevents addition of new headers\n to already large requests or to POST requests. It is generally wise to set it\n to about 1024. It is automatically readjusted to half of bufsize if it is\n larger than that. This means you don't have to worry about it when changing\n bufsize.\n\ntune.rcvbuf.client <number>\ntune.rcvbuf.server <number>\n Forces the kernel socket receive buffer size on the client or the server side\n to the specified value in bytes. This value applies to all TCP/HTTP frontends\n and backends. It should normally never be set, and the default size (0) lets\n the kernel autotune this value depending on the amount of available memory.\n However it can sometimes help to set it to very low values (eg: 4096) in\n order to save kernel memory by preventing it from buffering too large amounts\n of received data. Lower values will significantly increase CPU usage though.\n\ntune.sndbuf.client <number>\ntune.sndbuf.server <number>\n Forces the kernel socket send buffer size on the client or the server side to\n the specified value in bytes. This value applies to all TCP/HTTP frontends\n and backends. It should normally never be set, and the default size (0) lets\n the kernel autotune this value depending on the amount of available memory.\n However it can sometimes help to set it to very low values (eg: 4096) in\n order to save kernel memory by preventing it from buffering too large amounts\n of received data. Lower values will significantly increase CPU usage though.\n Another use case is to prevent write timeouts with extremely slow clients due\n to the kernel waiting for a large part of the buffer to be read before\n notifying haproxy again.\n\n\n3.3. Debugging\n--------------\n\ndebug\n Enables debug mode which dumps to stdout all exchanges, and disables forking\n into background. It is the equivalent of the command-line argument "-d". It\n should never be used in a production configuration since it may prevent full\n system startup.\n\nquiet\n Do not display any message during startup. It is equivalent to the command-\n line argument "-q".\n\n3.4. Userlists\n--------------\nIt is possible to control access to frontend/backend/listen sections or to\nhttp stats by allowing only authenticated and authorized users. To do this,\nit is required to create at least one userlist and to define users.\n\nuserlist <listname>\n Creates new userlist with name <listname>. Many independent userlists can be\n used to store authentication & authorization data for independent customers.\n\ngroup <groupname> [users <user>,<user>,(...)]\n Adds group <groupname> to the current userlist. It is also possible to\n attach users to this group by using a comma separated list of names\n proceeded by "users" keyword.\n\nuser <username> [password|insecure-password <password>]\n [groups <group>,<group>,(...)]\n Adds user <username> to the current userlist. Both secure (encrypted) and\n insecure (unencrypted) passwords can be used. Encrypted passwords are\n evaluated using the crypt(3) function so depending of the system's\n capabilities, different algorithms are supported. For example modern Glibc\n based Linux system supports MD5, SHA-256, SHA-512 and of course classic,\n DES-based method of crypting passwords.\n\n\n Example:\n userlist L1\n group G1 users tiger,scott\n group G2 users xdb,scott\n\n user tiger password $6$k6y3o.eP$JlKBx9za9667qe4(...)xHSwRv6J.C0/D7cV91\n user scott insecure-password elgato\n user xdb insecure-password hello\n\n userlist L2\n group G1\n group G2\n\n user tiger password $6$k6y3o.eP$JlKBx(...)xHSwRv6J.C0/D7cV91 groups G1\n user scott insecure-password elgato groups G1,G2\n user xdb insecure-password hello groups G2\n\n Please note that both lists are functionally identical.\n\n4. Proxies\n----------\n\nProxy configuration can be located in a set of sections :\n - defaults <name>\n - frontend <name>\n - backend <name>\n - listen <name>\n\nA "defaults" section sets default parameters for all other sections following\nits declaration. Those default parameters are reset by the next "defaults"\nsection. See below for the list of parameters which can be set in a "defaults"\nsection. The name is optional but its use is encouraged for better readability.\n\nA "frontend" section describes a set of listening sockets accepting client\nconnections.\n\nA "backend" section describes a set of servers to which the proxy will connect\nto forward incoming connections.\n\nA "listen" section defines a complete proxy with its frontend and backend\nparts combined in one section. It is generally useful for TCP-only traffic.\n\nAll proxy names must be formed from upper and lower case letters, digits,\n'-' (dash), '_' (underscore) , '.' (dot) and ':' (colon). ACL names are\ncase-sensitive, which means that "www" and "WWW" are two different proxies.\n\nHistorically, all proxy names could overlap, it just caused troubles in the\nlogs. Since the introduction of content switching, it is mandatory that two\nproxies with overlapping capabilities (frontend/backend) have different names.\nHowever, it is still permitted that a frontend and a backend share the same\nname, as this configuration seems to be commonly encountered.\n\nRight now, two major proxy modes are supported : "tcp", also known as layer 4,\nand "http", also known as layer 7. In layer 4 mode, HAProxy simply forwards\nbidirectional traffic between two sides. In layer 7 mode, HAProxy analyzes the\nprotocol, and can interact with it by allowing, blocking, switching, adding,\nmodifying, or removing arbitrary contents in requests or responses, based on\narbitrary criteria.\n\n\n4.1. Proxy keywords matrix\n--------------------------\n\nThe following list of keywords is supported. Most of them may only be used in a\nlimited set of section types. Some of them are marked as "deprecated" because\nthey are inherited from an old syntax which may be confusing or functionally\nlimited, and there are new recommended keywords to replace them. Keywords\nmarked with "(*)" can be optionally inverted using the "no" prefix, eg. "no\noption contstats". This makes sense when the option has been enabled by default\nand must be disabled for a specific instance. Such options may also be prefixed\nwith "default" in order to restore default settings regardless of what has been\nspecified in a previous "defaults" section.\n\n\n keyword defaults frontend listen backend\n------------------------------------+----------+----------+---------+---------\nacl - X X X\nappsession - - X X\nbacklog X X X -\nbalance X - X X\nbind - X X -\nbind-process X X X X\nblock - X X X\ncapture cookie - X X -\ncapture request header - X X -\ncapture response header - X X -\nclitimeout (deprecated) X X X -\ncontimeout (deprecated) X - X X\ncookie X - X X\ndefault-server X - X X\ndefault_backend X X X -\ndescription - X X X\ndisabled X X X X\ndispatch - - X X\nenabled X X X X\nerrorfile X X X X\nerrorloc X X X X\nerrorloc302 X X X X\n-- keyword -------------------------- defaults - frontend - listen -- backend -\nerrorloc303 X X X X\nforce-persist - X X X\nfullconn X - X X\ngrace X X X X\nhash-type X - X X\nhttp-check disable-on-404 X - X X\nhttp-check expect - - X X\nhttp-check send-state X - X X\nhttp-request - X X X\nid - X X X\nignore-persist - X X X\nlog X X X X\nmaxconn X X X -\nmode X X X X\nmonitor fail - X X -\nmonitor-net X X X -\nmonitor-uri X X X -\noption abortonclose (*) X - X X\noption accept-invalid-http-request (*) X X X -\noption accept-invalid-http-response (*) X - X X\noption allbackups (*) X - X X\noption checkcache (*) X - X X\noption clitcpka (*) X X X -\noption contstats (*) X X X -\noption dontlog-normal (*) X X X -\noption dontlognull (*) X X X -\noption forceclose (*) X X X X\n-- keyword -------------------------- defaults - frontend - listen -- backend -\noption forwardfor X X X X\noption http-no-delay (*) X X X X\noption http-pretend-keepalive (*) X X X X\noption http-server-close (*) X X X X\noption http-use-proxy-header (*) X X X -\noption httpchk X - X X\noption httpclose (*) X X X X\noption httplog X X X X\noption http_proxy (*) X X X X\noption independant-streams (*) X X X X\noption ldap-check X - X X\noption log-health-checks (*) X - X X\noption log-separate-errors (*) X X X -\noption logasap (*) X X X -\noption mysql-check X - X X\noption nolinger (*) X X X X\noption originalto X X X X\noption persist (*) X - X X\noption redispatch (*) X - X X\noption smtpchk X - X X\noption socket-stats (*) X X X -\noption splice-auto (*) X X X X\noption splice-request (*) X X X X\noption splice-response (*) X X X X\noption srvtcpka (*) X - X X\noption ssl-hello-chk X - X X\n-- keyword -------------------------- defaults - frontend - listen -- backend -\noption tcp-smart-accept (*) X X X -\noption tcp-smart-connect (*) X - X X\noption tcpka X X X X\noption tcplog X X X X\noption transparent (*) X - X X\npersist rdp-cookie X - X X\nrate-limit sessions X X X -\nredirect - X X X\nredisp (deprecated) X - X X\nredispatch (deprecated) X - X X\nreqadd - X X X\nreqallow - X X X\nreqdel - X X X\nreqdeny - X X X\nreqiallow - X X X\nreqidel - X X X\nreqideny - X X X\nreqipass - X X X\nreqirep - X X X\nreqisetbe - X X X\nreqitarpit - X X X\nreqpass - X X X\nreqrep - X X X\n-- keyword -------------------------- defaults - frontend - listen -- backend -\nreqsetbe - X X X\nreqtarpit - X X X\nretries X - X X\nrspadd - X X X\nrspdel - X X X\nrspdeny - X X X\nrspidel - X X X\nrspideny - X X X\nrspirep - X X X\nrsprep - X X X\nserver - - X X\nsource X - X X\nsrvtimeout (deprecated) X - X X\nstats admin - - X X\nstats auth X - X X\nstats enable X - X X\nstats hide-version X - X X\nstats http-request - - X X\nstats realm X - X X\nstats refresh X - X X\nstats scope X - X X\nstats show-desc X - X X\nstats show-legends X - X X\nstats show-node X - X X\nstats uri X - X X\n-- keyword -------------------------- defaults - frontend - listen -- backend -\nstick match - - X X\nstick on - - X X\nstick store-request - - X X\nstick-table - - X X\ntcp-request content accept - X X -\ntcp-request content reject - X X -\ntcp-request inspect-delay - X X -\ntimeout check X - X X\ntimeout client X X X -\ntimeout clitimeout (deprecated) X X X -\ntimeout connect X - X X\ntimeout contimeout (deprecated) X - X X\ntimeout http-keep-alive X X X X\ntimeout http-request X X X X\ntimeout queue X - X X\ntimeout server X - X X\ntimeout srvtimeout (deprecated) X - X X\ntimeout tarpit X X X X\ntransparent (deprecated) X - X X\nuse_backend - X X -\n------------------------------------+----------+----------+---------+---------\n keyword defaults frontend listen backend\n\n\n4.2. Alphabetically sorted keywords reference\n---------------------------------------------\n\nThis section provides a description of each keyword and its usage.\n\n\nacl <aclname> <criterion> [flags] [operator] <value> ...\n Declare or complete an access list.\n May be used in sections : defaults | frontend | listen | backend\n no | yes | yes | yes\n Example:\n acl invalid_src src 0.0.0.0/7 224.0.0.0/3\n acl invalid_src src_port 0:1023\n acl local_dst hdr(host) -i localhost\n\n See section 7 about ACL usage.\n\n\nappsession <cookie> len <length> timeout <holdtime>\n [request-learn] [prefix] [mode <path-parameters|query-string>]\n Define session stickiness on an existing application cookie.\n May be used in sections : defaults | frontend | listen | backend\n no | no | yes | yes\n Arguments :\n <cookie> this is the name of the cookie used by the application and which\n HAProxy will have to learn for each new session.\n\n <length> this is the max number of characters that will be memorized and\n checked in each cookie value.\n\n <holdtime> this is the time after which the cookie will be removed from\n memory if unused. If no unit is specified, this time is in\n milliseconds.\n\n request-learn\n If this option is specified, then haproxy will be able to learn\n the cookie found in the request in case the server does not\n specify any in response. This is typically what happens with\n PHPSESSID cookies, or when haproxy's session expires before\n the application's session and the correct server is selected.\n It is recommended to specify this option to improve reliability.\n\n prefix When this option is specified, haproxy will match on the cookie\n prefix (or URL parameter prefix). The appsession value is the\n data following this prefix.\n\n Example :\n appsession ASPSESSIONID len 64 timeout 3h prefix\n\n This will match the cookie ASPSESSIONIDXXXX=XXXXX,\n the appsession value will be XXXX=XXXXX.\n\n mode This option allows to change the URL parser mode.\n 2 modes are currently supported :\n - path-parameters :\n The parser looks for the appsession in the path parameters\n part (each parameter is separated by a semi-colon), which is\n convenient for JSESSIONID for example.\n This is the default mode if the option is not set.\n - query-string :\n In this mode, the parser will look for the appsession in the\n query string.\n\n When an application cookie is defined in a backend, HAProxy will check when\n the server sets such a cookie, and will store its value in a table, and\n associate it with the server's identifier. Up to <length> characters from\n the value will be retained. On each connection, haproxy will look for this\n cookie both in the "Cookie:" headers, and as a URL parameter (depending on\n the mode used). If a known value is found, the client will be directed to the\n server associated with this value. Otherwise, the load balancing algorithm is\n applied. Cookies are automatically removed from memory when they have been\n unused for a duration longer than <holdtime>.\n\n The definition of an application cookie is limited to one per backend.\n\n Note : Consider not using this feature in multi-process mode (nbproc > 1)\n unless you know what you do : memory is not shared between the\n processes, which can result in random behaviours.\n\n Example :\n appsession JSESSIONID len 52 timeout 3h\n\n See also : "cookie", "capture cookie", "balance", "stick", "stick-table",\n "ignore-persist", "nbproc" and "bind-process".\n\n\nbacklog <conns>\n Give hints to the system about the approximate listen backlog desired size\n May be used in sections : defaults | frontend | listen | backend\n yes | yes | yes | no\n Arguments :\n <conns> is the number of pending connections. Depending on the operating\n system, it may represent the number of already acknowledged\n connections, of non-acknowledged ones, or both.\n\n In order to protect against SYN flood attacks, one solution is to increase\n the system's SYN backlog size. Depending on the system, sometimes it is just\n tunable via a system parameter, sometimes it is not adjustable at all, and\n sometimes the system relies on hints given by the application at the time of\n the listen() syscall. By default, HAProxy passes the frontend's maxconn value\n to the listen() syscall. On systems which can make use of this value, it can\n sometimes be useful to be able to specify a different value, hence this\n backlog parameter.\n\n On Linux 2.4, the parameter is ignored by the system. On Linux 2.6, it is\n used as a hint and the system accepts up to the smallest greater power of\n two, and never more than some limits (usually 32768).\n\n See also : "maxconn" and the target operating system's tuning guide.\n\n\nbalance <algorithm> [ <arguments> ]\nbalance url_param <param> [check_post [<max_wait>]]\n Define the load balancing algorithm to be used in a backend.\n May be used in sections : defaults | frontend | listen | backend\n yes | no | yes | yes\n Arguments :\n <algorithm> is the algorithm used to select a server when doing load\n balancing. This only applies when no persistence information\n is available, or when a connection is redispatched to another\n server. <algorithm> may be one of the following :\n\n roundrobin Each server is used in turns, according to their weights.\n This is the smoothest and fairest algorithm when the server's\n processing time remains equally distributed. This algorithm\n is dynamic, which means that server weights may be adjusted\n on the fly for slow starts for instance. It is limited by\n design to 4128 active servers per backend. Note that in some\n large farms, when a server becomes up after having been down\n for a very short time, it may sometimes take a few hundreds\n requests for it to be re-integrated into the farm and start\n receiving traffic. This is normal, though very rare. It is\n indicated here in case you would have the chance to observe\n it, so that you don't worry.\n\n static-rr Each server is used in turns, according to their weights.\n This algorithm is as similar to roundrobin except that it is\n static, which means that changing a server's weight on the\n fly will have no effect. On the other hand, it has no design\n limitation on the number of servers, and when a server goes\n up, it is always immediately reintroduced into the farm, once\n the full map is recomputed. It also uses slightly less CPU to\n run (around -1%).\n\n leastconn The server with the lowest number of connections receives the\n connection. Round-robin is performed within groups of servers\n of the same load to ensure that all servers will be used. Use\n of this algorithm is recommended where very long sessions are\n expected, such as LDAP, SQL, TSE, etc... but is not very well\n suited for protocols using short sessions such as HTTP. This\n algorithm is dynamic, which means that server weights may be\n adjusted on the fly for slow starts for instance.\n\n source The source IP address is hashed and divided by the total\n weight of the running servers to designate which server will\n receive the request. This ensures that the same client IP\n address will always reach the same server as long as no\n server goes down or up. If the hash result changes due to the\n number of running servers changing, many clients will be\n directed to a different server. This algorithm is generally\n used in TCP mode where no cookie may be inserted. It may also\n be used on the Internet to provide a best-effort stickiness\n to clients which refuse session cookies. This algorithm is\n static by default, which means that changing a server's\n weight on the fly will have no effect, but this can be\n changed using "hash-type".\n\n uri This algorithm hashes either the left part of the URI (before\n the question mark) or the whole URI (if the "whole" parameter\n is present) and divides the hash value by the total weight of\n the running servers. The result designates which server will\n receive the request. This ensures that the same URI will\n always be directed to the same server as long as no server\n goes up or down. This is used with proxy caches and\n anti-virus proxies in order to maximize the cache hit rate.\n Note that this algorithm may only be used in an HTTP backend.\n This algorithm is static by default, which means that\n changing a server's weight on the fly will have no effect,\n but this can be changed using "hash-type".\n\n This algorithm supports two optional parameters "len" and\n "depth", both followed by a positive integer number. These\n options may be helpful when it is needed to balance servers\n based on the beginning of the URI only. The "len" parameter\n indicates that the algorithm should only consider that many\n characters at the beginning of the URI to compute the hash.\n Note that having "len" set to 1 rarely makes sense since most\n URIs start with a leading "/".\n\n The "depth" parameter indicates the maximum directory depth\n to be used to compute the hash. One level is counted for each\n slash in the request. If both parameters are specified, the\n evaluation stops when either is reached.\n\n url_param The URL parameter specified in argument will be looked up in\n the query string of each HTTP GET request.\n\n If the modifier "check_post" is used, then an HTTP POST\n request entity will be searched for the parameter argument,\n when it is not found in a query string after a question mark\n ('?') in the URL. Optionally, specify a number of octets to\n wait for before attempting to search the message body. If the\n entity can not be searched, then round robin is used for each\n request. For instance, if your clients always send the LB\n parameter in the first 128 bytes, then specify that. The\n default is 48. The entity data will not be scanned until the\n required number of octets have arrived at the gateway, this\n is the minimum of: (default/max_wait, Content-Length or first\n chunk length). If Content-Length is missing or zero, it does\n not need to wait for more data than the client promised to\n send. When Content-Length is present and larger than\n <max_wait>, then waiting is limited to <max_wait> and it is\n assumed that this will be enough data to search for the\n presence of the parameter. In the unlikely event that\n Transfer-Encoding: chunked is used, only the first chunk is\n scanned. Parameter values separated by a chunk boundary, may\n be randomly balanced if at all.\n\n If the parameter is found followed by an equal sign ('=') and\n a value, then the value is hashed and divided by the total\n weight of the running servers. The result designates which\n server will receive the request.\n\n This is used to track user identifiers in requests and ensure\n that a same user ID will always be sent to the same server as\n long as no server goes up or down. If no value is found or if\n the parameter is not found, then a round robin algorithm is\n applied. Note that this algorithm may only be used in an HTTP\n backend. This algorithm is static by default, which means\n that changing a server's weight on the fly will have no\n effect, but this can be changed using "hash-type".\n\n hdr(<name>) The HTTP header <name> will be looked up in each HTTP request.\n Just as with the equivalent ACL 'hdr()' function, the header\n name in parenthesis is not case sensitive. If the header is\n absent or if it does not contain any value, the roundrobin\n algorithm is applied instead.\n\n An optional 'use_domain_only' parameter is available, for\n reducing the hash algorithm to the main domain part with some\n specific headers such as 'Host'. For instance, in the Host\n value "haproxy.1wt.eu", only "1wt" will be considered.\n\n This algorithm is static by default, which means that\n changing a server's weight on the fly will have no effect,\n but this can be changed using "hash-type".\n\n rdp-cookie\n rdp-cookie(name)\n The RDP cookie <name> (or "mstshash" if omitted) will be\n looked up and hashed for each incoming TCP request. Just as\n with the equivalent ACL 'req_rdp_cookie()' function, the name\n is not case-sensitive. This mechanism is useful as a degraded\n persistence mode, as it makes it possible to always send the\n same user (or the same session ID) to the same server. If the\n cookie is not found, the normal roundrobin algorithm is\n used instead.\n\n Note that for this to work, the frontend must ensure that an\n RDP cookie is already present in the request buffer. For this\n you must use 'tcp-request content accept' rule combined with\n a 'req_rdp_cookie_cnt' ACL.\n\n This algorithm is static by default, which means that\n changing a server's weight on the fly will have no effect,\n but this can be changed using "hash-type".\n\n <arguments> is an optional list of arguments which may be needed by some\n algorithms. Right now, only "url_param" and "uri" support an\n optional argument.\n\n balance uri [len <len>] [depth <depth>]\n balance url_param <param> [check_post [<max_wait>]]\n\n The load balancing algorithm of a backend is set to roundrobin when no other\n algorithm, mode nor option have been set. The algorithm may only be set once\n for each backend.\n\n Examples :\n balance roundrobin\n balance url_param userid\n balance url_param session_id check_post 64\n balance hdr(User-Agent)\n balance hdr(host)\n balance hdr(Host) use_domain_only\n\n Note: the following caveats and limitations on using the "check_post"\n extension with "url_param" must be considered :\n\n - all POST requests are eligible for consideration, because there is no way\n to determine if the parameters will be found in the body or entity which\n may contain binary data. Therefore another method may be required to\n restrict consideration of POST requests that have no URL parameters in\n the body. (see acl reqideny http_end)\n\n - using a <max_wait> value larger than the request buffer size does not\n make sense and is useless. The buffer size is set at build time, and\n defaults to 16 kB.\n\n - Content-Encoding is not supported, the parameter search will probably\n fail; and load balancing will fall back to Round Robin.\n\n - Expect: 100-continue is not supported, load balancing will fall back to\n Round Robin.\n\n - Transfer-Encoding (RFC2616 3.6.1) is only supported in the first chunk.\n If the entire parameter value is not present in the first chunk, the\n selection of server is undefined (actually, defined by how little\n actually appeared in the first chunk).\n\n - This feature does not support generation of a 100, 411 or 501 response.\n\n - In some cases, requesting "check_post" MAY attempt to scan the entire\n contents of a message body. Scanning normally terminates when linear\n white space or control characters are found, indicating the end of what\n might be a URL parameter list. This is probably not a concern with SGML\n type message bodies.\n\n See also : "dispatch", "cookie", "appsession", "transparent", "hash-type" and\n "http_proxy".\n\n\nbind [<address>]:<port_range> [, ...]\nbind [<address>]:<port_range> [, ...] interface <interface>\nbind [<address>]:<port_range> [, ...] mss <maxseg>\nbind [<address>]:<port_range> [, ...] transparent\nbind [<address>]:<port_range> [, ...] id <id>\nbind [<address>]:<port_range> [, ...] name <name>\nbind [<address>]:<port_range> [, ...] defer-accept\n Define one or several listening addresses and/or ports in a frontend.\n May be used in sections : defaults | frontend | listen | backend\n no | yes | yes | no\n Arguments :\n <address> is optional and can be a host name, an IPv4 address, an IPv6\n address, or '*'. It designates the address the frontend will\n listen on. If unset, all IPv4 addresses of the system will be\n listened on. The same will apply for '*' or the system's\n special address "0.0.0.0".\n\n <port_range> is either a unique TCP port, or a port range for which the\n proxy will accept connections for the IP address specified\n above. The port is mandatory. Note that in the case of an\n IPv6 address, the port is always the number after the last\n colon (':'). A range can either be :\n - a numerical port (ex: '80')\n - a dash-delimited ports range explicitly stating the lower\n and upper bounds (ex: '2000-2100') which are included in\n the range.\n\n Particular care must be taken against port ranges, because\n every <address:port> couple consumes one socket (= a file\n descriptor), so it's easy to consume lots of descriptors\n with a simple range, and to run out of sockets. Also, each\n <address:port> couple must be used only once among all\n instances running on a same system. Please note that binding\n to ports lower than 1024 generally require particular\n privileges to start the program, which are independant of\n the 'uid' parameter.\n\n <interface> is an optional physical interface name. This is currently\n only supported on Linux. The interface must be a physical\n interface, not an aliased interface. When specified, all\n addresses on the same line will only be accepted if the\n incoming packet physically come through the designated\n interface. It is also possible to bind multiple frontends to\n the same address if they are bound to different interfaces.\n Note that binding to a physical interface requires root\n privileges.\n\n <maxseg> is an optional TCP Maximum Segment Size (MSS) value to be\n advertised on incoming connections. This can be used to force\n a lower MSS for certain specific ports, for instance for\n connections passing through a VPN. Note that this relies on a\n kernel feature which is theorically supported under Linux but\n was buggy in all versions prior to 2.6.28. It may or may not\n work on other operating systems. The commonly advertised\n value on Ethernet networks is 1460 = 1500(MTU) - 40(IP+TCP).\n\n <id> is a persistent value for socket ID. Must be positive and\n unique in the proxy. An unused value will automatically be\n assigned if unset. Can only be used when defining only a\n single socket.\n\n <name> is an optional name provided for stats\n\n transparent is an optional keyword which is supported only on certain\n Linux kernels. It indicates that the addresses will be bound\n even if they do not belong to the local machine. Any packet\n targeting any of these addresses will be caught just as if\n the address was locally configured. This normally requires\n that IP forwarding is enabled. Caution! do not use this with\n the default address '*', as it would redirect any traffic for\n the specified port. This keyword is available only when\n HAProxy is built with USE_LINUX_TPROXY=1.\n\n defer-accept is an optional keyword which is supported only on certain\n Linux kernels. It states that a connection will only be\n accepted once some data arrive on it, or at worst after the\n first retransmit. This should be used only on protocols for\n which the client talks first (eg: HTTP). It can slightly\n improve performance by ensuring that most of the request is\n already available when the connection is accepted. On the\n other hand, it will not be able to detect connections which\n don't talk. It is important to note that this option is\n broken in all kernels up to 2.6.31, as the connection is\n never accepted until the client talks. This can cause issues\n with front firewalls which would see an established\n connection while the proxy will only see it in SYN_RECV.\n\n It is possible to specify a list of address:port combinations delimited by\n commas. The frontend will then listen on all of these addresses. There is no\n fixed limit to the number of addresses and ports which can be listened on in\n a frontend, as well as there is no limit to the number of "bind" statements\n in a frontend.\n\n Example :\n listen http_proxy\n bind :80,:443\n bind 10.0.0.1:10080,10.0.0.1:10443\n\n See also : "source".\n\n\nbind-process [ all | odd | even | <number 1-32> ] ...\n Limit visibility of an instance to a certain set of processes numbers.\n May be used in sections : defaults | frontend | listen | backend\n yes | yes | yes | yes\n Arguments :\n all All process will see this instance. This is the default. It\n may be used to override a default value.\n\n odd This instance will be enabled on processes 1,3,5,...31. This\n option may be combined with other numbers.\n\n even This instance will be enabled on processes 2,4,6,...32. This\n option may be combined with other numbers. Do not use it\n with less than 2 processes otherwise some instances might be\n missing from all processes.\n\n number The instance will be enabled on this process number, between\n 1 and 32. You must be careful not to reference a process\n number greater than the configured global.nbproc, otherwise\n some instances might be missing from all processes.\n\n This keyword limits binding of certain instances to certain processes. This\n is useful in order not to have too many processes listening to the same\n ports. For instance, on a dual-core machine, it might make sense to set\n 'nbproc 2' in the global section, then distributes the listeners among 'odd'\n and 'even' instances.\n\n At the moment, it is not possible to reference more than 32 processes using\n this keyword, but this should be more than enough for most setups. Please\n note that 'all' really means all processes and is not limited to the first\n 32.\n\n If some backends are referenced by frontends bound to other processes, the\n backend automatically inherits the frontend's processes.\n\n Example :\n listen app_ip1\n bind 10.0.0.1:80\n bind-process odd\n\n listen app_ip2\n bind 10.0.0.2:80\n bind-process even\n\n listen management\n bind 10.0.0.3:80\n bind-process 1 2 3 4\n\n See also : "nbproc" in global section.\n\n\nblock { if | unless } <condition>\n Block a layer 7 request if/unless a condition is matched\n May be used in sections : defaults | frontend | listen | backend\n no | yes | yes | yes\n\n The HTTP request will be blocked very early in the layer 7 processing\n if/unless <condition> is matched. A 403 error will be returned if the request\n is blocked. The condition has to reference ACLs (see section 7). This is\n typically used to deny access to certain sensitive resources if some\n conditions are met or not met. There is no fixed limit to the number of\n "block" statements per instance.\n\n Example:\n acl invalid_src src 0.0.0.0/7 224.0.0.0/3\n acl invalid_src src_port 0:1023\n acl local_dst hdr(host) -i localhost\n block if invalid_src || local_dst\n\n See section 7 about ACL usage.\n\n\ncapture cookie <name> len <length>\n Capture and log a cookie in the request and in the response.\n May be used in sections : defaults | frontend | listen | backend\n no | yes | yes | no\n Arguments :\n <name> is the beginning of the name of the cookie to capture. In order\n to match the exact name, simply suffix the name with an equal\n sign ('='). The full name will appear in the logs, which is\n useful with application servers which adjust both the cookie name\n and value (eg: ASPSESSIONXXXXX).\n\n <length> is the maximum number of characters to report in the logs, which\n include the cookie name, the equal sign and the value, all in the\n standard "name=value" form. The string will be truncated on the\n right if it exceeds <length>.\n\n Only the first cookie is captured. Both the "cookie" request headers and the\n "set-cookie" response headers are monitored. This is particularly useful to\n check for application bugs causing session crossing or stealing between\n users, because generally the user's cookies can only change on a login page.\n\n When the cookie was not presented by the client, the associated log column\n will report "-". When a request does not cause a cookie to be assigned by the\n server, a "-" is reported in the response column.\n\n The capture is performed in the frontend only because it is necessary that\n the log format does not change for a given frontend depending on the\n backends. This may change in the future. Note that there can be only one\n "capture cookie" statement in a frontend. The maximum capture length is\n configured in the sources by default to 64 characters. It is not possible to\n specify a capture in a "defaults" section.\n\n Example:\n capture cookie ASPSESSION len 32\n\n See also : "capture request header", "capture response header" as well as\n section 8 about logging.\n\n\ncapture request header <name> len <length>\n Capture and log the first occurrence of the specified request header.\n May be used in sections : defaults | frontend | listen | backend\n no | yes | yes | no\n Arguments :\n <name> is the name of the header to capture. The header names are not\n case-sensitive, but it is a common practice to write them as they\n appear in the requests, with the first letter of each word in\n upper case. The header name will not appear in the logs, only the\n value is reported, but the position in the logs is respected.\n\n <length> is the maximum number of characters to extract from the value and\n report in the logs. The string will be truncated on the right if\n it exceeds <length>.\n\n Only the first value of the last occurrence of the header is captured. The\n value will be added to the logs between braces ('{}'). If multiple headers\n are captured, they will be delimited by a vertical bar ('|') and will appear\n in the same order they were declared in the configuration. Non-existent\n headers will be logged just as an empty string. Common uses for request\n header captures include the "Host" field in virtual hosting environments, the\n "Content-length" when uploads are supported, "User-agent" to quickly\n differentiate between real users and robots, and "X-Forwarded-For" in proxied\n environments to find where the request came from.\n\n Note that when capturing headers such as "User-agent", some spaces may be\n logged, making the log analysis more difficult. Thus be careful about what\n you log if you know your log parser is not smart enough to rely on the\n braces.\n\n There is no limit to the number of captured request headers, but each capture\n is limited to 64 characters. In order to keep log format consistent for a\n same frontend, header captures can only be declared in a frontend. It is not\n possible to specify a capture in a "defaults" section.\n\n Example:\n capture request header Host len 15\n capture request header X-Forwarded-For len 15\n capture request header Referrer len 15\n\n See also : "capture cookie", "capture response header" as well as section 8\n about logging.\n\n\ncapture response header <name> len <length>\n Capture and log the first occurrence of the specified response header.\n May be used in sections : defaults | frontend | listen | backend\n no | yes | yes | no\n Arguments :\n <name> is the name of the header to capture. The header names are not\n case-sensitive, but it is a common practice to write them as they\n appear in the response, with the first letter of each word in\n upper case. The header name will not appear in the logs, only the\n value is reported, but the position in the logs is respected.\n\n <length> is the maximum number of characters to extract from the value and\n report in the logs. The string will be truncated on the right if\n it exceeds <length>.\n\n Only the first value of the last occurrence of the header is captured. The\n result will be added to the logs between braces ('{}') after the captured\n request headers. If multiple headers are captured, they will be delimited by\n a vertical bar ('|') and will appear in the same order they were declared in\n the configuration. Non-existent headers will be logged just as an empty\n string. Common uses for response header captures include the "Content-length"\n header which indicates how many bytes are expected to be returned, the\n "Location" header to track redirections.\n\n There is no limit to the number of captured response headers, but each\n capture is limited to 64 characters. In order to keep log format consistent\n for a same frontend, header captures can only be declared in a frontend. It\n is not possible to specify a capture in a "defaults" section.\n\n Example:\n capture response header Content-length len 9\n capture response header Location len 15\n\n See also : "capture cookie", "capture request header" as well as section 8\n about logging.\n\n\nclitimeout <timeout> (deprecated)\n Set the maximum inactivity time on the client side.\n May be used in sections : defaults | frontend | listen | backend\n yes | yes | yes | no\n Arguments :\n <timeout> is the timeout value is specified in milliseconds by default, but\n can be in any other unit if the number is suffixed by the unit,\n as explained at the top of this document.\n\n The inactivity timeout applies when the client is expected to acknowledge or\n send data. In HTTP mode, this timeout is particularly important to consider\n during the first phase, when the client sends the request, and during the\n response while it is reading data sent by the server. The value is specified\n in milliseconds by default, but can be in any other unit if the number is\n suffixed by the unit, as specified at the top of this document. In TCP mode\n (and to a lesser extent, in HTTP mode), it is highly recommended that the\n client timeout remains equal to the server timeout in order to avoid complex\n situations to debug. It is a good practice to cover one or several TCP packet\n losses by specifying timeouts that are slightly above multiples of 3 seconds\n (eg: 4 or 5 seconds).\n\n This parameter is specific to frontends, but can be specified once for all in\n "defaults" sections. This is in fact one of the easiest solutions not to\n forget about it. An unspecified timeout results in an infinite timeout, which\n is not recommended. Such a usage is accepted and works but reports a warning\n during startup because it may results in accumulation of expired sessions in\n the system if the system's timeouts are not configured either.\n\n This parameter is provided for compatibility but is currently deprecated.\n Please use "timeout client" instead.\n\n See also : "timeout client", "timeout http-request", "timeout server", and\n "srvtimeout".\n\n\ncontimeout <timeout> (deprecated)\n Set the maximum time to wait for a connection attempt to a server to succeed.\n May be used in sections : defaults | frontend | listen | backend\n yes | no | yes | yes\n Arguments :\n <timeout> is the timeout value is specified in milliseconds by default, but\n can be in any other unit if the number is suffixed by the unit,\n as explained at the top of this document.\n\n If the server is located on the same LAN as haproxy, the connection should be\n immediate (less than a few milliseconds). Anyway, it is a good practice to\n cover one or several TCP packet losses by specifying timeouts that are\n slightly above multiples of 3 seconds (eg: 4 or 5 seconds). By default, the\n connect timeout also presets the queue timeout to the same value if this one\n has not been specified. Historically, the contimeout was also used to set the\n tarpit timeout in a listen section, which is not possible in a pure frontend.\n\n This parameter is specific to backends, but can be specified once for all in\n "defaults" sections. This is in fact one of the easiest solutions not to\n forget about it. An unspecified timeout results in an infinite timeout, which\n is not recommended. Such a usage is accepted and works but reports a warning\n during startup because it may results in accumulation of failed sessions in\n the system if the system's timeouts are not configured either.\n\n This parameter is provided for backwards compatibility but is currently\n deprecated. Please use "timeout connect", "timeout queue" or "timeout tarpit"\n instead.\n\n See also : "timeout connect", "timeout queue", "timeout tarpit",\n "timeout server", "contimeout".\n\n\ncookie <name> [ rewrite | insert | prefix ] [ indirect ] [ nocache ]\n [ postonly ] [ preserve ] [ httponly ] [ secure ]\n [ domain <domain> ]* [ maxidle <idle> ] [ maxlife <life> ]\n Enable cookie-based persistence in a backend.\n May be used in sections : defaults | frontend | listen | backend\n yes | no | yes | yes\n Arguments :\n <name> is the name of the cookie which will be monitored, modified or\n inserted in order to bring persistence. This cookie is sent to\n the client via a "Set-Cookie" header in the response, and is\n brought back by the client in a "Cookie" header in all requests.\n Special care should be taken to choose a name which does not\n conflict with any likely application cookie. Also, if the same\n backends are subject to be used by the same clients (eg:\n HTTP/HTTPS), care should be taken to use different cookie names\n between all backends if persistence between them is not desired.\n\n rewrite This keyword indicates that the cookie will be provided by the\n server and that haproxy will have to modify its value to set the\n server's identifier in it. This mode is handy when the management\n of complex combinations of "Set-cookie" and "Cache-control"\n headers is left to the application. The application can then\n decide whether or not it is appropriate to emit a persistence\n cookie. Since all responses should be monitored, this mode only\n works in HTTP close mode. Unless the application behaviour is\n very complex and/or broken, it is advised not to start with this\n mode for new deployments. This keyword is incompatible with\n "insert" and "prefix".\n\n insert This keyword indicates that the persistence cookie will have to\n be inserted by haproxy in server responses if the client did not\n\n already have a cookie that would have permitted it to access this\n server. When used without the "preserve" option, if the server\n emits a cookie with the same name, it will be remove before\n processing. For this reason, this mode can be used to upgrade\n existing configurations running in the "rewrite" mode. The cookie\n will only be a session cookie and will not be stored on the\n client's disk. By default, unless the "indirect" option is added,\n the server will see the cookies emitted by the client. Due to\n caching effects, it is generally wise to add the "nocache" or\n "postonly" keywords (see below). The "insert" keyword is not\n compatible with "rewrite" and "prefix".\n\n prefix This keyword indicates that instead of relying on a dedicated\n cookie for the persistence, an existing one will be completed.\n This may be needed in some specific environments where the client\n does not support more than one single cookie and the application\n already needs it. In this case, whenever the server sets a cookie\n named <name>, it will be prefixed with the server's identifier\n and a delimiter. The prefix will be removed from all client\n requests so that the server still finds the cookie it emitted.\n Since all requests and responses are subject to being modified,\n this mode requires the HTTP close mode. The "prefix" keyword is\n not compatible with "rewrite" and "insert".\n\n indirect When this option is specified, no cookie will be emitted to a\n client which already has a valid one for the server which has\n processed the request. If the server sets such a cookie itself,\n it will be removed, unless the "preserve" option is also set. In\n "insert" mode, this will additionally remove cookies from the\n requests transmitted to the server, making the persistence\n mechanism totally transparent from an application point of view.\n\n nocache This option is recommended in conjunction with the insert mode\n when there is a cache between the client and HAProxy, as it\n ensures that a cacheable response will be tagged non-cacheable if\n a cookie needs to be inserted. This is important because if all\n persistence cookies are added on a cacheable home page for\n instance, then all customers will then fetch the page from an\n outer cache and will all share the same persistence cookie,\n leading to one server receiving much more traffic than others.\n See also the "insert" and "postonly" options.\n\n postonly This option ensures that cookie insertion will only be performed\n on responses to POST requests. It is an alternative to the\n "nocache" option, because POST responses are not cacheable, so\n this ensures that the persistence cookie will never get cached.\n Since most sites do not need any sort of persistence before the\n first POST which generally is a login request, this is a very\n efficient method to optimize caching without risking to find a\n persistence cookie in the cache.\n See also the "insert" and "nocache" options.\n\n preserve This option may only be used with "insert" and/or "indirect". It\n allows the server to emit the persistence cookie itself. In this\n case, if a cookie is found in the response, haproxy will leave it\n untouched. This is useful in order to end persistence after a\n logout request for instance. For this, the server just has to\n emit a cookie with an invalid value (eg: empty) or with a date in\n the past. By combining this mechanism with the "disable-on-404"\n check option, it is possible to perform a completely graceful\n shutdown because users will definitely leave the server after\n they logout.\n\n httponly This option tells haproxy to add an "HttpOnly" cookie attribute\n when a cookie is inserted. This attribute is used so that a\n user agent doesn't share the cookie with non-HTTP components.\n Please check RFC6265 for more information on this attribute.\n\n secure This option tells haproxy to add a "Secure" cookie attribute when\n a cookie is inserted. This attribute is used so that a user agent\n never emits this cookie over non-secure channels, which means\n that a cookie learned with this flag will be presented only over\n SSL/TLS connections. Please check RFC6265 for more information on\n this attribute.\n\n domain This option allows to specify the domain at which a cookie is\n inserted. It requires exactly one parameter: a valid domain\n name. If the domain begins with a dot, the browser is allowed to\n use it for any host ending with that name. It is also possible to\n specify several domain names by invoking this option multiple\n times. Some browsers might have small limits on the number of\n domains, so be careful when doing that. For the record, sending\n 10 domains to MSIE 6 or Firefox 2 works as expected.\n\n maxidle This option allows inserted cookies to be ignored after some idle\n time. It only works with insert-mode cookies. When a cookie is\n sent to the client, the date this cookie was emitted is sent too.\n Upon further presentations of this cookie, if the date is older\n than the delay indicated by the parameter (in seconds), it will\n be ignored. Otherwise, it will be refreshed if needed when the\n response is sent to the client. This is particularly useful to\n prevent users who never close their browsers from remaining for\n too long on the same server (eg: after a farm size change). When\n this option is set and a cookie has no date, it is always\n accepted, but gets refreshed in the response. This maintains the\n ability for admins to access their sites. Cookies that have a\n date in the future further than 24 hours are ignored. Doing so\n lets admins fix timezone issues without risking kicking users off\n the site.\n\n maxlife This option allows inserted cookies to be ignored after some life\n time, whether they're in use or not. It only works with insert\n mode cookies. When a cookie is first sent to the client, the date\n this cookie was emitted is sent too. Upon further presentations\n of this cookie, if the date is older than the delay indicated by\n the parameter (in seconds), it will be ignored. If the cookie in\n the request has no date, it is accepted and a date will be set.\n Cookies that have a date in the future further than 24 hours are\n ignored. Doing so lets admins fix timezone issues without risking\n kicking users off the site. Contrary to maxidle, this value is\n not refreshed, only the first visit date counts. Both maxidle and\n maxlife may be used at the time. This is particularly useful to\n prevent users who never close their browsers from remaining for\n too long on the same server (eg: after a farm size change). This\n is stronger than the maxidle method in that it forces a\n redispatch after some absolute delay.\n\n There can be only one persistence cookie per HTTP backend, and it can be\n declared in a defaults section. The value of the cookie will be the value\n indicated after the "cookie" keyword in a "server" statement. If no cookie\n is declared for a given server, the cookie is not set.\n\n Examples :\n cookie JSESSIONID prefix\n cookie SRV insert indirect nocache\n cookie SRV insert postonly indirect\n cookie SRV insert indirect nocache maxidle 30m maxlife 8h\n\n See also : "appsession", "balance source", "capture cookie", "server"\n and "ignore-persist".\n\n\ndefault-server [param*]\n Change default options for a server in a backend\n May be used in sections : defaults | frontend | listen | backend\n yes | no | yes | yes\n Arguments:\n <param*> is a list of parameters for this server. The "default-server"\n keyword accepts an important number of options and has a complete\n section dedicated to it. Please refer to section 5 for more\n details.\n\n Example :\n default-server inter 1000 weight 13\n\n See also: "server" and section 5 about server options\n\n\ndefault_backend <backend>\n Specify the backend to use when no "use_backend" rule has been matched.\n May be used in sections : defaults | frontend | listen | backend\n yes | yes | yes | no\n Arguments :\n <backend> is the name of the backend to use.\n\n When doing content-switching between frontend and backends using the\n "use_backend" keyword, it is often useful to indicate which backend will be\n used when no rule has matched. It generally is the dynamic backend which\n will catch all undetermined requests.\n\n Example :\n\n use_backend dynamic if url_dyn\n use_backend static if url_css url_img extension_img\n default_backend dynamic\n\n See also : "use_backend", "reqsetbe", "reqisetbe"\n\n\ndisabled\n Disable a proxy, frontend or backend.\n May be used in sections : defaults | frontend | listen | backend\n yes | yes | yes | yes\n Arguments :\n\n The "disabled" keyword is used to disable an instance, mainly in order to\n liberate a listening port or to temporarily disable a service. The instance\n will still be created and its configuration will be checked, but it will be\n created in the "stopped" state and will appear as such in the statistics. It\n will not receive any traffic nor will it send any health-checks or logs. It\n is possible to disable many instances at once by adding the "disabled"\n keyword in a "defaults" section.\n\n See also : "enabled"\n\n\ndispatch <address>:<port>\n Set a default server address\n May be used in sections : defaults | frontend | listen | backend\n no | no | yes | yes\n Arguments : none\n\n <address> is the IPv4 address of the default server. Alternatively, a\n resolvable hostname is supported, but this name will be resolved\n during start-up.\n\n <ports> is a mandatory port specification. All connections will be sent\n to this port, and it is not permitted to use port offsets as is\n possible with normal servers.\n\n The "dispatch" keyword designates a default server for use when no other\n server can take the connection. In the past it was used to forward non\n persistent connections to an auxiliary load balancer. Due to its simple\n syntax, it has also been used for simple TCP relays. It is recommended not to\n use it for more clarity, and to use the "server" directive instead.\n\n See also : "server"\n\n\nenabled\n Enable a proxy, frontend or backend.\n May be used in sections : defaults | frontend | listen | backend\n yes | yes | yes | yes\n Arguments : none\n\n The "enabled" keyword is used to explicitly enable an instance, when the\n defaults has been set to "disabled". This is very rarely used.\n\n See also : "disabled"\n\n\nerrorfile <code> <file>\n Return a file contents instead of errors generated by HAProxy\n May be used in sections : defaults | frontend | listen | backend\n yes | yes | yes | yes\n Arguments :\n <code> is the HTTP status code. Currently, HAProxy is capable of\n generating codes 200, 400, 403, 408, 500, 502, 503, and 504.\n\n <file> designates a file containing the full HTTP response. It is\n recommended to follow the common practice of appending ".http" to\n the filename so that people do not confuse the response with HTML\n error pages, and to use absolute paths, since files are read\n before any chroot is performed.\n\n It is important to understand that this keyword is not meant to rewrite\n errors returned by the server, but errors detected and returned by HAProxy.\n This is why the list of supported errors is limited to a small set.\n\n Code 200 is emitted in response to requests matching a "monitor-uri" rule.\n\n The files are returned verbatim on the TCP socket. This allows any trick such\n as redirections to another URL or site, as well as tricks to clean cookies,\n force enable or disable caching, etc... The package provides default error\n files returning the same contents as default errors.\n\n The files should not exceed the configured buffer size (BUFSIZE), which\n generally is 8 or 16 kB, otherwise they will be truncated. It is also wise\n not to put any reference to local contents (eg: images) in order to avoid\n loops between the client and HAProxy when all servers are down, causing an\n error to be returned instead of an image. For better HTTP compliance, it is\n recommended that all header lines end with CR-LF and not LF alone.\n\n The files are read at the same time as the configuration and kept in memory.\n For this reason, the errors continue to be returned even when the process is\n chrooted, and no file change is considered while the process is running. A\n simple method for developing those files consists in associating them to the\n 403 status code and interrogating a blocked URL.\n\n See also : "errorloc", "errorloc302", "errorloc303"\n\n Example :\n errorfile 400 /etc/haproxy/errorfiles/400badreq.http\n errorfile 403 /etc/haproxy/errorfiles/403forbid.http\n errorfile 503 /etc/haproxy/errorfiles/503sorry.http\n\n\nerrorloc <code> <url>\nerrorloc302 <code> <url>\n Return an HTTP redirection to a URL instead of errors generated by HAProxy\n May be used in sections : defaults | frontend | listen | backend\n yes | yes | yes | yes\n Arguments :\n <code> is the HTTP status code. Currently, HAProxy is capable of\n generating codes 200, 400, 403, 408, 500, 502, 503, and 504.\n\n <url> it is the exact contents of the "Location" header. It may contain\n either a relative URI to an error page hosted on the same site,\n or an absolute URI designating an error page on another site.\n Special care should be given to relative URIs to avoid redirect\n loops if the URI itself may generate the same error (eg: 500).\n\n It is important to understand that this keyword is not meant to rewrite\n errors returned by the server, but errors detected and returned by HAProxy.\n This is why the list of supported errors is limited to a small set.\n\n Code 200 is emitted in response to requests matching a "monitor-uri" rule.\n\n Note that both keyword return the HTTP 302 status code, which tells the\n client to fetch the designated URL using the same HTTP method. This can be\n quite problematic in case of non-GET methods such as POST, because the URL\n sent to the client might not be allowed for something other than GET. To\n workaround this problem, please use "errorloc303" which send the HTTP 303\n status code, indicating to the client that the URL must be fetched with a GET\n request.\n\n See also : "errorfile", "errorloc303"\n\n\nerrorloc303 <code> <url>\n Return an HTTP redirection to a URL instead of errors generated by HAProxy\n May be used in sections : defaults | frontend | listen | backend\n yes | yes | yes | yes\n Arguments :\n <code> is the HTTP status code. Currently, HAProxy is capable of\n generating codes 400, 403, 408, 500, 502, 503, and 504.\n\n <url> it is the exact contents of the "Location" header. It may contain\n either a relative URI to an error page hosted on the same site,\n or an absolute URI designating an error page on another site.\n Special care should be given to relative URIs to avoid redirect\n loops if the URI itself may generate the same error (eg: 500).\n\n It is important to understand that this keyword is not meant to rewrite\n errors returned by the server, but errors detected and returned by HAProxy.\n This is why the list of supported errors is limited to a small set.\n\n Code 200 is emitted in response to requests matching a "monitor-uri" rule.\n\n Note that both keyword return the HTTP 303 status code, which tells the\n client to fetch the designated URL using the same HTTP GET method. This\n solves the usual problems associated with "errorloc" and the 302 code. It is\n possible that some very old browsers designed before HTTP/1.1 do not support\n it, but no such problem has been reported till now.\n\n See also : "errorfile", "errorloc", "errorloc302"\n\n\nforce-persist { if | unless } <condition>\n Declare a condition to force persistence on down servers\n May be used in sections: defaults | frontend | listen | backend\n no | yes | yes | yes\n\n By default, requests are not dispatched to down servers. It is possible to\n force this using "option persist", but it is unconditional and redispatches\n to a valid server if "option redispatch" is set. That leaves with very little\n possibilities to force some requests to reach a server which is artificially\n marked down for maintenance operations.\n\n The "force-persist" statement allows one to declare various ACL-based\n conditions which, when met, will cause a request to ignore the down status of\n a server and still try to connect to it. That makes it possible to start a\n server, still replying an error to the health checks, and run a specially\n configured browser to test the service. Among the handy methods, one could\n use a specific source IP address, or a specific cookie. The cookie also has\n the advantage that it can easily be added/removed on the browser from a test\n page. Once the service is validated, it is then possible to open the service\n to the world by returning a valid response to health checks.\n\n The forced persistence is enabled when an "if" condition is met, or unless an\n "unless" condition is met. The final redispatch is always disabled when this\n is used.\n\n See also : "option redispatch", "ignore-persist", "persist",\n and section 7 about ACL usage.\n\n\nfullconn <conns>\n Specify at what backend load the servers will reach their maxconn\n May be used in sections : defaults | frontend | listen | backend\n yes | no | yes | yes\n Arguments :\n <conns> is the number of connections on the backend which will make the\n servers use the maximal number of connections.\n\n When a server has a "maxconn" parameter specified, it means that its number\n of concurrent connections will never go higher. Additionally, if it has a\n "minconn" parameter, it indicates a dynamic limit following the backend's\n load. The server will then always accept at least <minconn> connections,\n never more than <maxconn>, and the limit will be on the ramp between both\n values when the backend has less than <conns> concurrent connections. This\n makes it possible to limit the load on the servers during normal loads, but\n push it further for important loads without overloading the servers during\n exceptional loads.\n\n Example :\n # The servers will accept between 100 and 1000 concurrent connections each\n # and the maximum of 1000 will be reached when the backend reaches 10000\n # connections.\n backend dynamic\n fullconn 10000\n server srv1 dyn1:80 minconn 100 maxconn 1000\n server srv2 dyn2:80 minconn 100 maxconn 1000\n\n See also : "maxconn", "server"\n\n\ngrace <time>\n Maintain a proxy operational for some time after a soft stop\n May be used in sections : defaults | frontend | listen | backend\n yes | yes | yes | yes\n Arguments :\n <time> is the time (by default in milliseconds) for which the instance\n will remain operational with the frontend sockets still listening\n when a soft-stop is received via the SIGUSR1 signal.\n\n This may be used to ensure that the services disappear in a certain order.\n This was designed so that frontends which are dedicated to monitoring by an\n external equipment fail immediately while other ones remain up for the time\n needed by the equipment to detect the failure.\n\n Note that currently, there is very little benefit in using this parameter,\n and it may in fact complicate the soft-reconfiguration process more than\n simplify it.\n\n\nhash-type <method>\n Specify a method to use for mapping hashes to servers\n May be used in sections : defaults | frontend | listen | backend\n yes | no | yes | yes\n Arguments :\n map-based the hash table is a static array containing all alive servers.\n The hashes will be very smooth, will consider weights, but will\n be static in that weight changes while a server is up will be\n ignored. This means that there will be no slow start. Also,\n since a server is selected by its position in the array, most\n mappings are changed when the server count changes. This means\n that when a server goes up or down, or when a server is added\n to a farm, most connections will be redistributed to different\n servers. This can be inconvenient with caches for instance.\n\n consistent the hash table is a tree filled with many occurrences of each\n server. The hash key is looked up in the tree and the closest\n server is chosen. This hash is dynamic, it supports changing\n weights while the servers are up, so it is compatible with the\n slow start feature. It has the advantage that when a server\n goes up or down, only its associations are moved. When a server\n is added to the farm, only a few part of the mappings are\n redistributed, making it an ideal algorithm for caches.\n However, due to its principle, the algorithm will never be very\n smooth and it may sometimes be necessary to adjust a server's\n weight or its ID to get a more balanced distribution. In order\n to get the same distribution on multiple load balancers, it is\n important that all servers have the same IDs.\n\n The default hash type is "map-based" and is recommended for most usages.\n\n See also : "balance", "server"\n\n\nhttp-check disable-on-404\n Enable a maintenance mode upon HTTP/404 response to health-checks\n May be used in sections : defaults | frontend | listen | backend\n yes | no | yes | yes\n Arguments : none\n\n When this option is set, a server which returns an HTTP code 404 will be\n excluded from further load-balancing, but will still receive persistent\n connections. This provides a very convenient method for Web administrators\n to perform a graceful shutdown of their servers. It is also important to note\n that a server which is detected as failed while it was in this mode will not\n generate an alert, just a notice. If the server responds 2xx or 3xx again, it\n will immediately be reinserted into the farm. The status on the stats page\n reports "NOLB" for a server in this mode. It is important to note that this\n option only works in conjunction with the "httpchk" option. If this option\n is used with "http-check expect", then it has precedence over it so that 404\n responses will still be considered as soft-stop.\n\n See also : "option httpchk", "http-check expect"\n\n\nhttp-check expect [!] <match> <pattern>\n Make HTTP health checks consider reponse contents or specific status codes\n May be used in sections : defaults | frontend | listen | backend\n yes | no | yes | yes\n Arguments :\n <match> is a keyword indicating how to look for a specific pattern in the\n response. The keyword may be one of "status", "rstatus",\n "string", or "rstring". The keyword may be preceeded by an\n exclamation mark ("!") to negate the match. Spaces are allowed\n between the exclamation mark and the keyword. See below for more\n details on the supported keywords.\n\n <pattern> is the pattern to look for. It may be a string or a regular\n expression. If the pattern contains spaces, they must be escaped\n with the usual backslash ('\').\n\n By default, "option httpchk" considers that response statuses 2xx and 3xx\n are valid, and that others are invalid. When "http-check expect" is used,\n it defines what is considered valid or invalid. Only one "http-check"\n statement is supported in a backend. If a server fails to respond or times\n out, the check obviously fails. The available matches are :\n\n status <string> : test the exact string match for the HTTP status code.\n A health check respose will be considered valid if the\n response's status code is exactly this string. If the\n "status" keyword is prefixed with "!", then the response\n will be considered invalid if the status code matches.\n\n rstatus <regex> : test a regular expression for the HTTP status code.\n A health check respose will be considered valid if the\n response's status code matches the expression. If the\n "rstatus" keyword is prefixed with "!", then the response\n will be considered invalid if the status code matches.\n This is mostly used to check for multiple codes.\n\n string <string> : test the exact string match in the HTTP response body.\n A health check respose will be considered valid if the\n response's body contains this exact string. If the\n "string" keyword is prefixed with "!", then the response\n will be considered invalid if the body contains this\n string. This can be used to look for a mandatory word at\n the end of a dynamic page, or to detect a failure when a\n specific error appears on the check page (eg: a stack\n trace).\n\n rstring <regex> : test a regular expression on the HTTP response body.\n A health check respose will be considered valid if the\n response's body matches this expression. If the "rstring"\n keyword is prefixed with "!", then the response will be\n considered invalid if the body matches the expression.\n This can be used to look for a mandatory word at the end\n of a dynamic page, or to detect a failure when a specific\n error appears on the check page (eg: a stack trace).\n\n It is important to note that the responses will be limited to a certain size\n defined by the global "tune.chksize" option, which defaults to 16384 bytes.\n Thus, too large responses may not contain the mandatory pattern when using\n "string" or "rstring". If a large response is absolutely required, it is\n possible to change the default max size by setting the global variable.\n However, it is worth keeping in mind that parsing very large responses can\n waste some CPU cycles, especially when regular expressions are used, and that\n it is always better to focus the checks on smaller resources.\n\n Last, if "http-check expect" is combined with "http-check disable-on-404",\n then this last one has precedence when the server responds with 404.\n\n Examples :\n # only accept status 200 as valid\n http-check expect status 200\n\n # consider SQL errors as errors\n http-check expect ! string SQL\ Error\n\n # consider status 5xx only as errors\n http-check expect ! rstatus ^5\n\n # check that we have a correct hexadecimal tag before /html\n http-check expect rstring <!--tag:[0-9a-f]*</html>\n\n See also : "option httpchk", "http-check disable-on-404"\n\n\nhttp-check send-state\n Enable emission of a state header with HTTP health checks\n May be used in sections : defaults | frontend | listen | backend\n yes | no | yes | yes\n Arguments : none\n\n When this option is set, haproxy will systematically send a special header\n "X-Haproxy-Server-State" with a list of parameters indicating to each server\n how they are seen by haproxy. This can be used for instance when a server is\n manipulated without access to haproxy and the operator needs to know whether\n haproxy still sees it up or not, or if the server is the last one in a farm.\n\n The header is composed of fields delimited by semi-colons, the first of which\n is a word ("UP", "DOWN", "NOLB"), possibly followed by a number of valid\n checks on the total number before transition, just as appears in the stats\n interface. Next headers are in the form "<variable>=<value>", indicating in\n no specific order some values available in the stats interface :\n - a variable "name", containing the name of the backend followed by a slash\n ("/") then the name of the server. This can be used when a server is\n checked in multiple backends.\n\n - a variable "node" containing the name of the haproxy node, as set in the\n global "node" variable, otherwise the system's hostname if unspecified.\n\n - a variable "weight" indicating the weight of the server, a slash ("/")\n and the total weight of the farm (just counting usable servers). This\n helps to know if other servers are available to handle the load when this\n one fails.\n\n - a variable "scur" indicating the current number of concurrent connections\n on the server, followed by a slash ("/") then the total number of\n connections on all servers of the same backend.\n\n - a variable "qcur" indicating the current number of requests in the\n server's queue.\n\n Example of a header received by the application server :\n >>> X-Haproxy-Server-State: UP 2/3; name=bck/srv2; node=lb1; weight=1/2; \\n scur=13/22; qcur=0\n\n See also : "option httpchk", "http-check disable-on-404"\n\nhttp-request { allow | deny | auth [realm <realm>] }\n [ { if | unless } <condition> ]\n Access control for Layer 7 requests\n\n May be used in sections: defaults | frontend | listen | backend\n no | yes | yes | yes\n\n These set of options allow to fine control access to a\n frontend/listen/backend. Each option may be followed by if/unless and acl.\n First option with matched condition (or option without condition) is final.\n For "deny" a 403 error will be returned, for "allow" normal processing is\n performed, for "auth" a 401/407 error code is returned so the client\n should be asked to enter a username and password.\n\n There is no fixed limit to the number of http-request statements per\n instance.\n\n Example:\n acl nagios src 192.168.129.3\n acl local_net src 192.168.0.0/16\n acl auth_ok http_auth(L1)\n\n http-request allow if nagios\n http-request allow if local_net auth_ok\n http-request auth realm Gimme if local_net auth_ok\n http-request deny\n\n Example:\n acl auth_ok http_auth_group(L1) G1\n\n http-request auth unless auth_ok\n\n See also : "stats http-request", section 3.4 about userlists and section 7\n about ACL usage.\n\nhttp-send-name-header [<header>]\n Add the server name to a request. Use the header string given by <header>\n\n May be used in sections: defaults | frontend | listen | backend\n yes | no | yes | yes\n\n Arguments :\n\n <header> The header string to use to send the server name\n\n The "http-send-name-header" statement causes the name of the target\n server to be added to the headers of an HTTP request. The name\n is added with the header string proved.\n\n See also : "server"\n\nid <value>\n Set a persistent ID to a proxy.\n May be used in sections : defaults | frontend | listen | backend\n no | yes | yes | yes\n Arguments : none\n\n Set a persistent ID for the proxy. This ID must be unique and positive.\n An unused ID will automatically be assigned if unset. The first assigned\n value will be 1. This ID is currently only returned in statistics.\n\n\nignore-persist { if | unless } <condition>\n Declare a condition to ignore persistence\n May be used in sections: defaults | frontend | listen | backend\n no | yes | yes | yes\n\n By default, when cookie persistence is enabled, every requests containing\n the cookie are unconditionally persistent (assuming the target server is up\n and running).\n\n The "ignore-persist" statement allows one to declare various ACL-based\n conditions which, when met, will cause a request to ignore persistence.\n This is sometimes useful to load balance requests for static files, which\n oftenly don't require persistence. This can also be used to fully disable\n persistence for a specific User-Agent (for example, some web crawler bots).\n\n Combined with "appsession", it can also help reduce HAProxy memory usage, as\n the appsession table won't grow if persistence is ignored.\n\n The persistence is ignored when an "if" condition is met, or unless an\n "unless" condition is met.\n\n See also : "force-persist", "cookie", and section 7 about ACL usage.\n\n\nlog global\nlog <address> <facility> [<level> [<minlevel>]]\n Enable per-instance logging of events and traffic.\n May be used in sections : defaults | frontend | listen | backend\n yes | yes | yes | yes\n Arguments :\n global should be used when the instance's logging parameters are the\n same as the global ones. This is the most common usage. "global"\n replaces <address>, <facility> and <level> with those of the log\n entries found in the "global" section. Only one "log global"\n statement may be used per instance, and this form takes no other\n parameter.\n\n <address> indicates where to send the logs. It takes the same format as\n for the "global" section's logs, and can be one of :\n\n - An IPv4 address optionally followed by a colon (':') and a UDP\n port. If no port is specified, 514 is used by default (the\n standard syslog port).\n\n - A filesystem path to a UNIX domain socket, keeping in mind\n considerations for chroot (be sure the path is accessible\n inside the chroot) and uid/gid (be sure the path is\n appropriately writeable).\n\n <facility> must be one of the 24 standard syslog facilities :\n\n kern user mail daemon auth syslog lpr news\n uucp cron auth2 ftp ntp audit alert cron2\n local0 local1 local2 local3 local4 local5 local6 local7\n\n <level> is optional and can be specified to filter outgoing messages. By\n default, all messages are sent. If a level is specified, only\n messages with a severity at least as important as this level\n will be sent. An optional minimum level can be specified. If it\n is set, logs emitted with a more severe level than this one will\n be capped to this level. This is used to avoid sending "emerg"\n messages on all terminals on some default syslog configurations.\n Eight levels are known :\n\n emerg alert crit err warning notice info debug\n\n Note that up to two "log" entries may be specified per instance. However, if\n "log global" is used and if the "global" section already contains 2 log\n entries, then additional log entries will be ignored.\n\n Also, it is important to keep in mind that it is the frontend which decides\n what to log from a connection, and that in case of content switching, the log\n entries from the backend will be ignored. Connections are logged at level\n "info".\n\n However, backend log declaration define how and where servers status changes\n will be logged. Level "notice" will be used to indicate a server going up,\n "warning" will be used for termination signals and definitive service\n termination, and "alert" will be used for when a server goes down.\n\n Note : According to RFC3164, messages are truncated to 1024 bytes before\n being emitted.\n\n Example :\n log global\n log 127.0.0.1:514 local0 notice # only send important events\n log 127.0.0.1:514 local0 notice notice # same but limit output level\n\n\nmaxconn <conns>\n Fix the maximum number of concurrent connections on a frontend\n May be used in sections : defaults | frontend | listen | backend\n yes | yes | yes | no\n Arguments :\n <conns> is the maximum number of concurrent connections the frontend will\n accept to serve. Excess connections will be queued by the system\n in the socket's listen queue and will be served once a connection\n closes.\n\n If the system supports it, it can be useful on big sites to raise this limit\n very high so that haproxy manages connection queues, instead of leaving the\n clients with unanswered connection attempts. This value should not exceed the\n global maxconn. Also, keep in mind that a connection contains two buffers\n of 8kB each, as well as some other data resulting in about 17 kB of RAM being\n consumed per established connection. That means that a medium system equipped\n with 1GB of RAM can withstand around 40000-50000 concurrent connections if\n properly tuned.\n\n Also, when <conns> is set to large values, it is possible that the servers\n are not sized to accept such loads, and for this reason it is generally wise\n to assign them some reasonable connection limits.\n\n By default, this value is set to 2000.\n\n See also : "server", global section's "maxconn", "fullconn"\n\n\nmode { tcp|http|health }\n Set the running mode or protocol of the instance\n May be used in sections : defaults | frontend | listen | backend\n yes | yes | yes | yes\n Arguments :\n tcp The instance will work in pure TCP mode. A full-duplex connection\n will be established between clients and servers, and no layer 7\n examination will be performed. This is the default mode. It\n should be used for SSL, SSH, SMTP, ...\n\n http The instance will work in HTTP mode. The client request will be\n analyzed in depth before connecting to any server. Any request\n which is not RFC-compliant will be rejected. Layer 7 filtering,\n processing and switching will be possible. This is the mode which\n brings HAProxy most of its value.\n\n health The instance will work in "health" mode. It will just reply "OK"\n to incoming connections and close the connection. Nothing will be\n logged. This mode is used to reply to external components health\n checks. This mode is deprecated and should not be used anymore as\n it is possible to do the same and even better by combining TCP or\n HTTP modes with the "monitor" keyword.\n\n When doing content switching, it is mandatory that the frontend and the\n backend are in the same mode (generally HTTP), otherwise the configuration\n will be refused.\n\n Example :\n defaults http_instances\n mode http\n\n See also : "monitor", "monitor-net"\n\n\nmonitor fail { if | unless } <condition>\n Add a condition to report a failure to a monitor HTTP request.\n May be used in sections : defaults | frontend | listen | backend\n no | yes | yes | no\n Arguments :\n if <cond> the monitor request will fail if the condition is satisfied,\n and will succeed otherwise. The condition should describe a\n combined test which must induce a failure if all conditions\n are met, for instance a low number of servers both in a\n backend and its backup.\n\n unless <cond> the monitor request will succeed only if the condition is\n satisfied, and will fail otherwise. Such a condition may be\n based on a test on the presence of a minimum number of active\n servers in a list of backends.\n\n This statement adds a condition which can force the response to a monitor\n request to report a failure. By default, when an external component queries\n the URI dedicated to monitoring, a 200 response is returned. When one of the\n conditions above is met, haproxy will return 503 instead of 200. This is\n very useful to report a site failure to an external component which may base\n routing advertisements between multiple sites on the availability reported by\n haproxy. In this case, one would rely on an ACL involving the "nbsrv"\n criterion. Note that "monitor fail" only works in HTTP mode. Both status\n messages may be tweaked using "errorfile" or "errorloc" if needed.\n\n Example:\n frontend www\n mode http\n acl site_dead nbsrv(dynamic) lt 2\n acl site_dead nbsrv(static) lt 2\n monitor-uri /site_alive\n monitor fail if site_dead\n\n See also : "monitor-net", "monitor-uri", "errorfile", "errorloc"\n\n\nmonitor-net <source>\n Declare a source network which is limited to monitor requests\n May be used in sections : defaults | frontend | listen | backend\n yes | yes | yes | no\n Arguments :\n <source> is the source IPv4 address or network which will only be able to\n get monitor responses to any request. It can be either an IPv4\n address, a host name, or an address followed by a slash ('/')\n followed by a mask.\n\n In TCP mode, any connection coming from a source matching <source> will cause\n the connection to be immediately closed without any log. This allows another\n equipment to probe the port and verify that it is still listening, without\n forwarding the connection to a remote server.\n\n In HTTP mode, a connection coming from a source matching <source> will be\n accepted, the following response will be sent without waiting for a request,\n then the connection will be closed : "HTTP/1.0 200 OK". This is normally\n enough for any front-end HTTP probe to detect that the service is UP and\n running without forwarding the request to a backend server.\n\n Monitor requests are processed very early. It is not possible to block nor\n divert them using ACLs. They cannot be logged either, and it is the intended\n purpose. They are only used to report HAProxy's health to an upper component,\n nothing more. Right now, it is not possible to set failure conditions on\n requests caught by "monitor-net".\n\n Last, please note that only one "monitor-net" statement can be specified in\n a frontend. If more than one is found, only the last one will be considered.\n\n Example :\n # addresses .252 and .253 are just probing us.\n frontend www\n monitor-net 192.168.0.252/31\n\n See also : "monitor fail", "monitor-uri"\n\n\nmonitor-uri <uri>\n Intercept a URI used by external components' monitor requests\n May be used in sections : defaults | frontend | listen | backend\n yes | yes | yes | no\n Arguments :\n <uri> is the exact URI which we want to intercept to return HAProxy's\n health status instead of forwarding the request.\n\n When an HTTP request referencing <uri> will be received on a frontend,\n HAProxy will not forward it nor log it, but instead will return either\n "HTTP/1.0 200 OK" or "HTTP/1.0 503 Service unavailable", depending on failure\n conditions defined with "monitor fail". This is normally enough for any\n front-end HTTP probe to detect that the service is UP and running without\n forwarding the request to a backend server. Note that the HTTP method, the\n version and all headers are ignored, but the request must at least be valid\n at the HTTP level. This keyword may only be used with an HTTP-mode frontend.\n\n Monitor requests are processed very early. It is not possible to block nor\n divert them using ACLs. They cannot be logged either, and it is the intended\n purpose. They are only used to report HAProxy's health to an upper component,\n nothing more. However, it is possible to add any number of conditions using\n "monitor fail" and ACLs so that the result can be adjusted to whatever check\n can be imagined (most often the number of available servers in a backend).\n\n Example :\n # Use /haproxy_test to report haproxy's status\n frontend www\n mode http\n monitor-uri /haproxy_test\n\n See also : "monitor fail", "monitor-net"\n\n\noption abortonclose\nno option abortonclose\n Enable or disable early dropping of aborted requests pending in queues.\n May be used in sections : defaults | frontend | listen | backend\n yes | no | yes | yes\n Arguments : none\n\n In presence of very high loads, the servers will take some time to respond.\n The per-instance connection queue will inflate, and the response time will\n increase respective to the size of the queue times the average per-session\n response time. When clients will wait for more than a few seconds, they will\n often hit the "STOP" button on their browser, leaving a useless request in\n the queue, and slowing down other users, and the servers as well, because the\n request will eventually be served, then aborted at the first error\n encountered while delivering the response.\n\n As there is no way to distinguish between a full STOP and a simple output\n close on the client side, HTTP agents should be conservative and consider\n that the client might only have closed its output channel while waiting for\n the response. However, this introduces risks of congestion when lots of users\n do the same, and is completely useless nowadays because probably no client at\n all will close the session while waiting for the response. Some HTTP agents\n support this behaviour (Squid, Apache, HAProxy), and others do not (TUX, most\n hardware-based load balancers). So the probability for a closed input channel\n to represent a user hitting the "STOP" button is close to 100%, and the risk\n of being the single component to break rare but valid traffic is extremely\n low, which adds to the temptation to be able to abort a session early while\n still not served and not pollute the servers.\n\n In HAProxy, the user can choose the desired behaviour using the option\n "abortonclose". By default (without the option) the behaviour is HTTP\n compliant and aborted requests will be served. But when the option is\n specified, a session with an incoming channel closed will be aborted while\n it is still possible, either pending in the queue for a connection slot, or\n during the connection establishment if the server has not yet acknowledged\n the connection request. This considerably reduces the queue size and the load\n on saturated servers when users are tempted to click on STOP, which in turn\n reduces the response time for other users.\n\n If this option has been enabled in a "defaults" section, it can be disabled\n in a specific instance by prepending the "no" keyword before it.\n\n See also : "timeout queue" and server's "maxconn" and "maxqueue" parameters\n\n\noption accept-invalid-http-request\nno option accept-invalid-http-request\n Enable or disable relaxing of HTTP request parsing\n May be used in sections : defaults | frontend | listen | backend\n yes | yes | yes | no\n Arguments : none\n\n By default, HAProxy complies with RFC2616 in terms of message parsing. This\n means that invalid characters in header names are not permitted and cause an\n error to be returned to the client. This is the desired behaviour as such\n forbidden characters are essentially used to build attacks exploiting server\n weaknesses, and bypass security filtering. Sometimes, a buggy browser or\n server will emit invalid header names for whatever reason (configuration,\n implementation) and the issue will not be immediately fixed. In such a case,\n it is possible to relax HAProxy's header name parser to accept any character\n even if that does not make sense, by specifying this option.\n\n This option should never be enabled by default as it hides application bugs\n and open security breaches. It should only be deployed after a problem has\n been confirmed.\n\n When this option is enabled, erroneous header names will still be accepted in\n requests, but the complete request will be captured in order to permit later\n analysis using the "show errors" request on the UNIX stats socket. Doing this\n also helps confirming that the issue has been solved.\n\n If this option has been enabled in a "defaults" section, it can be disabled\n in a specific instance by prepending the "no" keyword before it.\n\n See also : "option accept-invalid-http-response" and "show errors" on the\n stats socket.\n\n\noption accept-invalid-http-response\nno option accept-invalid-http-response\n Enable or disable relaxing of HTTP response parsing\n May be used in sections : defaults | frontend | listen | backend\n yes | no | yes | yes\n Arguments : none\n\n By default, HAProxy complies with RFC2616 in terms of message parsing. This\n means that invalid characters in header names are not permitted and cause an\n error to be returned to the client. This is the desired behaviour as such\n forbidden characters are essentially used to build attacks exploiting server\n weaknesses, and bypass security filtering. Sometimes, a buggy browser or\n server will emit invalid header names for whatever reason (configuration,\n implementation) and the issue will not be immediately fixed. In such a case,\n it is possible to relax HAProxy's header name parser to accept any character\n even if that does not make sense, by specifying this option.\n\n This option should never be enabled by default as it hides application bugs\n and open security breaches. It should only be deployed after a problem has\n been confirmed.\n\n When this option is enabled, erroneous header names will still be accepted in\n responses, but the complete response will be captured in order to permit\n later analysis using the "show errors" request on the UNIX stats socket.\n Doing this also helps confirming that the issue has been solved.\n\n If this option has been enabled in a "defaults" section, it can be disabled\n in a specific instance by prepending the "no" keyword before it.\n\n See also : "option accept-invalid-http-request" and "show errors" on the\n stats socket.\n\n\noption allbackups\nno option allbackups\n Use either all backup servers at a time or only the first one\n May be used in sections : defaults | frontend | listen | backend\n yes | no | yes | yes\n Arguments : none\n\n By default, the first operational backup server gets all traffic when normal\n servers are all down. Sometimes, it may be preferred to use multiple backups\n at once, because one will not be enough. When "option allbackups" is enabled,\n the load balancing will be performed among all backup servers when all normal\n ones are unavailable. The same load balancing algorithm will be used and the\n servers' weights will be respected. Thus, there will not be any priority\n order between the backup servers anymore.\n\n This option is mostly used with static server farms dedicated to return a\n "sorry" page when an application is completely offline.\n\n If this option has been enabled in a "defaults" section, it can be disabled\n in a specific instance by prepending the "no" keyword before it.\n\n\noption checkcache\nno option checkcache\n Analyze all server responses and block requests with cacheable cookies\n May be used in sections : defaults | frontend | listen | backend\n yes | no | yes | yes\n Arguments : none\n\n Some high-level frameworks set application cookies everywhere and do not\n always let enough control to the developer to manage how the responses should\n be cached. When a session cookie is returned on a cacheable object, there is a\n high risk of session crossing or stealing between users traversing the same\n caches. In some situations, it is better to block the response than to let\n some sensitive session information go in the wild.\n\n The option "checkcache" enables deep inspection of all server responses for\n strict compliance with HTTP specification in terms of cacheability. It\n carefully checks "Cache-control", "Pragma" and "Set-cookie" headers in server\n response to check if there's a risk of caching a cookie on a client-side\n proxy. When this option is enabled, the only responses which can be delivered\n to the client are :\n - all those without "Set-Cookie" header ;\n - all those with a return code other than 200, 203, 206, 300, 301, 410,\n provided that the server has not set a "Cache-control: public" header ;\n - all those that come from a POST request, provided that the server has not\n set a 'Cache-Control: public' header ;\n - those with a 'Pragma: no-cache' header\n - those with a 'Cache-control: private' header\n - those with a 'Cache-control: no-store' header\n - those with a 'Cache-control: max-age=0' header\n - those with a 'Cache-control: s-maxage=0' header\n - those with a 'Cache-control: no-cache' header\n - those with a 'Cache-control: no-cache="set-cookie"' header\n - those with a 'Cache-control: no-cache="set-cookie,' header\n (allowing other fields after set-cookie)\n\n If a response doesn't respect these requirements, then it will be blocked\n just as if it was from an "rspdeny" filter, with an "HTTP 502 bad gateway".\n The session state shows "PH--" meaning that the proxy blocked the response\n during headers processing. Additionally, an alert will be sent in the logs so\n that admins are informed that there's something to be fixed.\n\n Due to the high impact on the application, the application should be tested\n in depth with the option enabled before going to production. It is also a\n good practice to always activate it during tests, even if it is not used in\n production, as it will report potentially dangerous application behaviours.\n\n If this option has been enabled in a "defaults" section, it can be disabled\n in a specific instance by prepending the "no" keyword before it.\n\n\noption clitcpka\nno option clitcpka\n Enable or disable the sending of TCP keepalive packets on the client side\n May be used in sections : defaults | frontend | listen | backend\n yes | yes | yes | no\n Arguments : none\n\n When there is a firewall or any session-aware component between a client and\n a server, and when the protocol involves very long sessions with long idle\n periods (eg: remote desktops), there is a risk that one of the intermediate\n components decides to expire a session which has remained idle for too long.\n\n Enabling socket-level TCP keep-alives makes the system regularly send packets\n to the other end of the connection, leaving it active. The delay between\n keep-alive probes is controlled by the system only and depends both on the\n operating system and its tuning parameters.\n\n It is important to understand that keep-alive packets are neither emitted nor\n received at the application level. It is only the network stacks which sees\n them. For this reason, even if one side of the proxy already uses keep-alives\n to maintain its connection alive, those keep-alive packets will not be\n forwarded to the other side of the proxy.\n\n Please note that this has nothing to do with HTTP keep-alive.\n\n Using option "clitcpka" enables the emission of TCP keep-alive probes on the\n client side of a connection, which should help when session expirations are\n noticed between HAProxy and a client.\n\n If this option has been enabled in a "defaults" section, it can be disabled\n in a specific instance by prepending the "no" keyword before it.\n\n See also : "option srvtcpka", "option tcpka"\n\n\noption contstats\n Enable continuous traffic statistics updates\n May be used in sections : defaults | frontend | listen | backend\n yes | yes | yes | no\n Arguments : none\n\n By default, counters used for statistics calculation are incremented\n only when a session finishes. It works quite well when serving small\n objects, but with big ones (for example large images or archives) or\n with A/V streaming, a graph generated from haproxy counters looks like\n a hedgehog. With this option enabled counters get incremented continuously,\n during a whole session. Recounting touches a hotpath directly so\n it is not enabled by default, as it has small performance impact (~0.5%).\n\n\noption dontlog-normal\nno option dontlog-normal\n Enable or disable logging of normal, successful connections\n May be used in sections : defaults | frontend | listen | backend\n yes | yes | yes | no\n Arguments : none\n\n There are large sites dealing with several thousand connections per second\n and for which logging is a major pain. Some of them are even forced to turn\n logs off and cannot debug production issues. Setting this option ensures that\n normal connections, those which experience no error, no timeout, no retry nor\n redispatch, will not be logged. This leaves disk space for anomalies. In HTTP\n mode, the response status code is checked and return codes 5xx will still be\n logged.\n\n It is strongly discouraged to use this option as most of the time, the key to\n complex issues is in the normal logs which will not be logged here. If you\n need to separate logs, see the "log-separate-errors" option instead.\n\n See also : "log", "dontlognull", "log-separate-errors" and section 8 about\n logging.\n\n\noption dontlognull\nno option dontlognull\n Enable or disable logging of null connections\n May be used in sections : defaults | frontend | listen | backend\n yes | yes | yes | no\n Arguments : none\n\n In certain environments, there are components which will regularly connect to\n various systems to ensure that they are still alive. It can be the case from\n another load balancer as well as from monitoring systems. By default, even a\n simple port probe or scan will produce a log. If those connections pollute\n the logs too much, it is possible to enable option "dontlognull" to indicate\n that a connection on which no data has been transferred will not be logged,\n which typically corresponds to those probes.\n\n It is generally recommended not to use this option in uncontrolled\n environments (eg: internet), otherwise scans and other malicious activities\n would not be logged.\n\n If this option has been enabled in a "defaults" section, it can be disabled\n in a specific instance by prepending the "no" keyword before it.\n\n See also : "log", "monitor-net", "monitor-uri" and section 8 about logging.\n\n\noption forceclose\nno option forceclose\n Enable or disable active connection closing after response is transferred.\n May be used in sections : defaults | frontend | listen | backend\n yes | yes | yes | yes\n Arguments : none\n\n Some HTTP servers do not necessarily close the connections when they receive\n the "Connection: close" set by "option httpclose", and if the client does not\n close either, then the connection remains open till the timeout expires. This\n causes high number of simultaneous connections on the servers and shows high\n global session times in the logs.\n\n When this happens, it is possible to use "option forceclose". It will\n actively close the outgoing server channel as soon as the server has finished\n to respond. This option implicitly enables the "httpclose" option. Note that\n this option also enables the parsing of the full request and response, which\n means we can close the connection to the server very quickly, releasing some\n resources earlier than with httpclose.\n\n This option may also be combined with "option http-pretend-keepalive", which\n will disable sending of the "Connection: close" header, but will still cause\n the connection to be closed once the whole response is received.\n\n If this option has been enabled in a "defaults" section, it can be disabled\n in a specific instance by prepending the "no" keyword before it.\n\n See also : "option httpclose" and "option http-pretend-keepalive"\n\n\noption forwardfor [ except <network> ] [ header <name> ] [ if-none ]\n Enable insertion of the X-Forwarded-For header to requests sent to servers\n May be used in sections : defaults | frontend | listen | backend\n yes | yes | yes | yes\n Arguments :\n <network> is an optional argument used to disable this option for sources\n matching <network>\n <name> an optional argument to specify a different "X-Forwarded-For"\n header name.\n\n Since HAProxy works in reverse-proxy mode, the servers see its IP address as\n their client address. This is sometimes annoying when the client's IP address\n is expected in server logs. To solve this problem, the well-known HTTP header\n "X-Forwarded-For" may be added by HAProxy to all requests sent to the server.\n This header contains a value representing the client's IP address. Since this\n header is always appended at the end of the existing header list, the server\n must be configured to always use the last occurrence of this header only. See\n the server's manual to find how to enable use of this standard header. Note\n that only the last occurrence of the header must be used, since it is really\n possible that the client has already brought one.\n\n The keyword "header" may be used to supply a different header name to replace\n the default "X-Forwarded-For". This can be useful where you might already\n have a "X-Forwarded-For" header from a different application (eg: stunnel),\n and you need preserve it. Also if your backend server doesn't use the\n "X-Forwarded-For" header and requires different one (eg: Zeus Web Servers\n require "X-Cluster-Client-IP").\n\n Sometimes, a same HAProxy instance may be shared between a direct client\n access and a reverse-proxy access (for instance when an SSL reverse-proxy is\n used to decrypt HTTPS traffic). It is possible to disable the addition of the\n header for a known source address or network by adding the "except" keyword\n followed by the network address. In this case, any source IP matching the\n network will not cause an addition of this header. Most common uses are with\n private networks or 127.0.0.1.\n\n Alternatively, the keyword "if-none" states that the header will only be\n added if it is not present. This should only be used in perfectly trusted\n environment, as this might cause a security issue if headers reaching haproxy\n are under the control of the end-user.\n\n This option may be specified either in the frontend or in the backend. If at\n least one of them uses it, the header will be added. Note that the backend's\n setting of the header subargument takes precedence over the frontend's if\n both are defined. In the case of the "if-none" argument, if at least one of\n the frontend or the backend does not specify it, it wants the addition to be\n mandatory, so it wins.\n\n It is important to note that by default, HAProxy works in tunnel mode and\n only inspects the first request of a connection, meaning that only the first\n request will have the header appended, which is certainly not what you want.\n In order to fix this, ensure that any of the "httpclose", "forceclose" or\n "http-server-close" options is set when using this option.\n\n Examples :\n # Public HTTP address also used by stunnel on the same machine\n frontend www\n mode http\n option forwardfor except 127.0.0.1 # stunnel already adds the header\n\n # Those servers want the IP Address in X-Client\n backend www\n mode http\n option forwardfor header X-Client\n\n See also : "option httpclose", "option http-server-close",\n "option forceclose"\n\n\noption http-no-delay\nno option http-no-delay\n Instruct the system to favor low interactive delays over performance in HTTP\n May be used in sections : defaults | frontend | listen | backend\n yes | yes | yes | yes\n Arguments : none\n\n In HTTP, each payload is unidirectional and has no notion of interactivity.\n Any agent is expected to queue data somewhat for a reasonably low delay.\n There are some very rare server-to-server applications that abuse the HTTP\n protocol and expect the payload phase to be highly interactive, with many\n interleaved data chunks in both directions within a single request. This is\n absolutely not supported by the HTTP specification and will not work across\n most proxies or servers. When such applications attempt to do this through\n haproxy, it works but they will experience high delays due to the network\n optimizations which favor performance by instructing the system to wait for\n enough data to be available in order to only send full packets. Typical\n delays are around 200 ms per round trip. Note that this only happens with\n abnormal uses. Normal uses such as CONNECT requests nor WebSockets are not\n affected.\n\n When "option http-no-delay" is present in either the frontend or the backend\n used by a connection, all such optimizations will be disabled in order to\n make the exchanges as fast as possible. Of course this offers no guarantee on\n the functionality, as it may break at any other place. But if it works via\n HAProxy, it will work as fast as possible. This option should never be used\n by default, and should never be used at all unless such a buggy application\n is discovered. The impact of using this option is an increase of bandwidth\n usage and CPU usage, which may significantly lower performance in high\n latency environments.\n\n\noption http-pretend-keepalive\nno option http-pretend-keepalive\n Define whether haproxy will announce keepalive to the server or not\n May be used in sections : defaults | frontend | listen | backend\n yes | yes | yes | yes\n Arguments : none\n\n When running with "option http-server-close" or "option forceclose", haproxy\n adds a "Connection: close" header to the request forwarded to the server.\n Unfortunately, when some servers see this header, they automatically refrain\n from using the chunked encoding for responses of unknown length, while this\n is totally unrelated. The immediate effect is that this prevents haproxy from\n maintaining the client connection alive. A second effect is that a client or\n a cache could receive an incomplete response without being aware of it, and\n consider the response complete.\n\n By setting "option http-pretend-keepalive", haproxy will make the server\n believe it will keep the connection alive. The server will then not fall back\n to the abnormal undesired above. When haproxy gets the whole response, it\n will close the connection with the server just as it would do with the\n "forceclose" option. That way the client gets a normal response and the\n connection is correctly closed on the server side.\n\n It is recommended not to enable this option by default, because most servers\n will more efficiently close the connection themselves after the last packet,\n and release its buffers slightly earlier. Also, the added packet on the\n network could slightly reduce the overall peak performance. However it is\n worth noting that when this option is enabled, haproxy will have slightly\n less work to do. So if haproxy is the bottleneck on the whole architecture,\n enabling this option might save a few CPU cycles.\n\n This option may be set both in a frontend and in a backend. It is enabled if\n at least one of the frontend or backend holding a connection has it enabled.\n This option may be compbined with "option httpclose", which will cause\n keepalive to be announced to the server and close to be announced to the\n client. This practice is discouraged though.\n\n If this option has been enabled in a "defaults" section, it can be disabled\n in a specific instance by prepending the "no" keyword before it.\n\n See also : "option forceclose" and "option http-server-close"\n\n\noption http-server-close\nno option http-server-close\n Enable or disable HTTP connection closing on the server side\n May be used in sections : defaults | frontend | listen | backend\n yes | yes | yes | yes\n Arguments : none\n\n By default, when a client communicates with a server, HAProxy will only\n analyze, log, and process the first request of each connection. Setting\n "option http-server-close" enables HTTP connection-close mode on the server\n side while keeping the ability to support HTTP keep-alive and pipelining on\n the client side. This provides the lowest latency on the client side (slow\n network) and the fastest session reuse on the server side to save server\n resources, similarly to "option forceclose". It also permits non-keepalive\n capable servers to be served in keep-alive mode to the clients if they\n conform to the requirements of RFC2616. Please note that some servers do not\n always conform to those requirements when they see "Connection: close" in the\n request. The effect will be that keep-alive will never be used. A workaround\n consists in enabling "option http-pretend-keepalive".\n\n At the moment, logs will not indicate whether requests came from the same\n session or not. The accept date reported in the logs corresponds to the end\n of the previous request, and the request time corresponds to the time spent\n waiting for a new request. The keep-alive request time is still bound to the\n timeout defined by "timeout http-keep-alive" or "timeout http-request" if\n not set.\n\n This option may be set both in a frontend and in a backend. It is enabled if\n at least one of the frontend or backend holding a connection has it enabled.\n It is worth noting that "option forceclose" has precedence over "option\n http-server-close" and that combining "http-server-close" with "httpclose"\n basically achieve the same result as "forceclose".\n\n If this option has been enabled in a "defaults" section, it can be disabled\n in a specific instance by prepending the "no" keyword before it.\n\n See also : "option forceclose", "option http-pretend-keepalive",\n "option httpclose" and "1.1. The HTTP transaction model".\n\n\noption http-use-proxy-header\nno option http-use-proxy-header\n Make use of non-standard Proxy-Connection header instead of Connection\n May be used in sections : defaults | frontend | listen | backend\n yes | yes | yes | no\n Arguments : none\n\n While RFC2616 explicitly states that HTTP/1.1 agents must use the\n Connection header to indicate their wish of persistent or non-persistent\n connections, both browsers and proxies ignore this header for proxied\n connections and make use of the undocumented, non-standard Proxy-Connection\n header instead. The issue begins when trying to put a load balancer between\n browsers and such proxies, because there will be a difference between what\n haproxy understands and what the client and the proxy agree on.\n\n By setting this option in a frontend, haproxy can automatically switch to use\n that non-standard header if it sees proxied requests. A proxied request is\n defined here as one where the URI begins with neither a '/' nor a '*'. The\n choice of header only affects requests passing through proxies making use of\n one of the "httpclose", "forceclose" and "http-server-close" options. Note\n that this option can only be specified in a frontend and will affect the\n request along its whole life.\n\n Also, when this option is set, a request which requires authentication will\n automatically switch to use proxy authentication headers if it is itself a\n proxied request. That makes it possible to check or enforce authentication in\n front of an existing proxy.\n\n This option should normally never be used, except in front of a proxy.\n\n See also : "option httpclose", "option forceclose" and "option\n http-server-close".\n\n\noption httpchk\noption httpchk <uri>\noption httpchk <method> <uri>\noption httpchk <method> <uri> <version>\n Enable HTTP protocol to check on the servers health\n May be used in sections : defaults | frontend | listen | backend\n yes | no | yes | yes\n Arguments :\n <method> is the optional HTTP method used with the requests. When not set,\n the "OPTIONS" method is used, as it generally requires low server\n processing and is easy to filter out from the logs. Any method\n may be used, though it is not recommended to invent non-standard\n ones.\n\n <uri> is the URI referenced in the HTTP requests. It defaults to " / "\n which is accessible by default on almost any server, but may be\n changed to any other URI. Query strings are permitted.\n\n <version> is the optional HTTP version string. It defaults to "HTTP/1.0"\n but some servers might behave incorrectly in HTTP 1.0, so turning\n it to HTTP/1.1 may sometimes help. Note that the Host field is\n mandatory in HTTP/1.1, and as a trick, it is possible to pass it\n after "\r\n" following the version string.\n\n By default, server health checks only consist in trying to establish a TCP\n connection. When "option httpchk" is specified, a complete HTTP request is\n sent once the TCP connection is established, and responses 2xx and 3xx are\n considered valid, while all other ones indicate a server failure, including\n the lack of any response.\n\n The port and interval are specified in the server configuration.\n\n This option does not necessarily require an HTTP backend, it also works with\n plain TCP backends. This is particularly useful to check simple scripts bound\n to some dedicated ports using the inetd daemon.\n\n Examples :\n # Relay HTTPS traffic to Apache instance and check service availability\n # using HTTP request "OPTIONS * HTTP/1.1" on port 80.\n backend https_relay\n mode tcp\n option httpchk OPTIONS * HTTP/1.1\r\nHost:\ www\n server apache1 192.168.1.1:443 check port 80\n\n See also : "option ssl-hello-chk", "option smtpchk", "option mysql-check",\n "http-check" and the "check", "port" and "inter" server options.\n\n\noption httpclose\nno option httpclose\n Enable or disable passive HTTP connection closing\n May be used in sections : defaults | frontend | listen | backend\n yes | yes | yes | yes\n Arguments : none\n\n By default, when a client communicates with a server, HAProxy will only\n analyze, log, and process the first request of each connection. If "option\n httpclose" is set, it will check if a "Connection: close" header is already\n set in each direction, and will add one if missing. Each end should react to\n this by actively closing the TCP connection after each transfer, thus\n resulting in a switch to the HTTP close mode. Any "Connection" header\n different from "close" will also be removed.\n\n It seldom happens that some servers incorrectly ignore this header and do not\n close the connection eventhough they reply "Connection: close". For this\n reason, they are not compatible with older HTTP 1.0 browsers. If this happens\n it is possible to use the "option forceclose" which actively closes the\n request connection once the server responds. Option "forceclose" also\n releases the server connection earlier because it does not have to wait for\n the client to acknowledge it.\n\n This option may be set both in a frontend and in a backend. It is enabled if\n at least one of the frontend or backend holding a connection has it enabled.\n If "option forceclose" is specified too, it has precedence over "httpclose".\n If "option http-server-close" is enabled at the same time as "httpclose", it\n basically achieves the same result as "option forceclose".\n\n If this option has been enabled in a "defaults" section, it can be disabled\n in a specific instance by prepending the "no" keyword before it.\n\n See also : "option forceclose", "option http-server-close" and\n "1.1. The HTTP transaction model".\n\n\noption httplog [ clf ]\n Enable logging of HTTP request, session state and timers\n May be used in sections : defaults | frontend | listen | backend\n yes | yes | yes | yes\n Arguments :\n clf if the "clf" argument is added, then the output format will be\n the CLF format instead of HAProxy's default HTTP format. You can\n use this when you need to feed HAProxy's logs through a specific\n log analyser which only support the CLF format and which is not\n extensible.\n\n By default, the log output format is very poor, as it only contains the\n source and destination addresses, and the instance name. By specifying\n "option httplog", each log line turns into a much richer format including,\n but not limited to, the HTTP request, the connection timers, the session\n status, the connections numbers, the captured headers and cookies, the\n frontend, backend and server name, and of course the source address and\n ports.\n\n This option may be set either in the frontend or the backend.\n\n If this option has been enabled in a "defaults" section, it can be disabled\n in a specific instance by prepending the "no" keyword before it. Specifying\n only "option httplog" will automatically clear the 'clf' mode if it was set\n by default.\n\n See also : section 8 about logging.\n\n\noption http_proxy\nno option http_proxy\n Enable or disable plain HTTP proxy mode\n May be used in sections : defaults | frontend | listen | backend\n yes | yes | yes | yes\n Arguments : none\n\n It sometimes happens that people need a pure HTTP proxy which understands\n basic proxy requests without caching nor any fancy feature. In this case,\n it may be worth setting up an HAProxy instance with the "option http_proxy"\n set. In this mode, no server is declared, and the connection is forwarded to\n the IP address and port found in the URL after the "http://" scheme.\n\n No host address resolution is performed, so this only works when pure IP\n addresses are passed. Since this option's usage perimeter is rather limited,\n it will probably be used only by experts who know they need exactly it. Last,\n if the clients are susceptible of sending keep-alive requests, it will be\n needed to add "option httpclose" to ensure that all requests will correctly\n be analyzed.\n\n If this option has been enabled in a "defaults" section, it can be disabled\n in a specific instance by prepending the "no" keyword before it.\n\n Example :\n # this backend understands HTTP proxy requests and forwards them directly.\n backend direct_forward\n option httpclose\n option http_proxy\n\n See also : "option httpclose"\n\n\noption independant-streams\nno option independant-streams\n Enable or disable independant timeout processing for both directions\n May be used in sections : defaults | frontend | listen | backend\n yes | yes | yes | yes\n Arguments : none\n\n By default, when data is sent over a socket, both the write timeout and the\n read timeout for that socket are refreshed, because we consider that there is\n activity on that socket, and we have no other means of guessing if we should\n receive data or not.\n\n While this default behaviour is desirable for almost all applications, there\n exists a situation where it is desirable to disable it, and only refresh the\n read timeout if there are incoming data. This happens on sessions with large\n timeouts and low amounts of exchanged data such as telnet session. If the\n server suddenly disappears, the output data accumulates in the system's\n socket buffers, both timeouts are correctly refreshed, and there is no way\n to know the server does not receive them, so we don't timeout. However, when\n the underlying protocol always echoes sent data, it would be enough by itself\n to detect the issue using the read timeout. Note that this problem does not\n happen with more verbose protocols because data won't accumulate long in the\n socket buffers.\n\n When this option is set on the frontend, it will disable read timeout updates\n on data sent to the client. There probably is little use of this case. When\n the option is set on the backend, it will disable read timeout updates on\n data sent to the server. Doing so will typically break large HTTP posts from\n slow lines, so use it with caution.\n\n See also : "timeout client" and "timeout server"\n\n\noption ldap-check\n Use LDAPv3 health checks for server testing\n May be used in sections : defaults | frontend | listen | backend\n yes | no | yes | yes\n Arguments : none\n\n It is possible to test that the server correctly talks LDAPv3 instead of just\n testing that it accepts the TCP connection. When this option is set, an\n LDAPv3 anonymous simple bind message is sent to the server, and the response\n is analyzed to find an LDAPv3 bind response message.\n\n The server is considered valid only when the LDAP response contains success\n resultCode (http://tools.ietf.org/html/rfc4511#section-4.1.9).\n\n Logging of bind requests is server dependent see your documentation how to\n configure it.\n\n Example :\n option ldap-check\n\n See also : "option httpchk"\n\n\noption log-health-checks\nno option log-health-checks\n Enable or disable logging of health checks\n May be used in sections : defaults | frontend | listen | backend\n yes | no | yes | yes\n Arguments : none\n\n Enable health checks logging so it possible to check for example what\n was happening before a server crash. Failed health check are logged if\n server is UP and succeeded health checks if server is DOWN, so the amount\n of additional information is limited.\n\n If health check logging is enabled no health check status is printed\n when servers is set up UP/DOWN/ENABLED/DISABLED.\n\n See also: "log" and section 8 about logging.\n\n\noption log-separate-errors\nno option log-separate-errors\n Change log level for non-completely successful connections\n May be used in sections : defaults | frontend | listen | backend\n yes | yes | yes | no\n Arguments : none\n\n Sometimes looking for errors in logs is not easy. This option makes haproxy\n raise the level of logs containing potentially interesting information such\n as errors, timeouts, retries, redispatches, or HTTP status codes 5xx. The\n level changes from "info" to "err". This makes it possible to log them\n separately to a different file with most syslog daemons. Be careful not to\n remove them from the original file, otherwise you would lose ordering which\n provides very important information.\n\n Using this option, large sites dealing with several thousand connections per\n second may log normal traffic to a rotating buffer and only archive smaller\n error logs.\n\n See also : "log", "dontlognull", "dontlog-normal" and section 8 about\n logging.\n\n\noption logasap\nno option logasap\n Enable or disable early logging of HTTP requests\n May be used in sections : defaults | frontend | listen | backend\n yes | yes | yes | no\n Arguments : none\n\n By default, HTTP requests are logged upon termination so that the total\n transfer time and the number of bytes appear in the logs. When large objects\n are being transferred, it may take a while before the request appears in the\n logs. Using "option logasap", the request gets logged as soon as the server\n sends the complete headers. The only missing information in the logs will be\n the total number of bytes which will indicate everything except the amount\n of data transferred, and the total time which will not take the transfer\n time into account. In such a situation, it's a good practice to capture the\n "Content-Length" response header so that the logs at least indicate how many\n bytes are expected to be transferred.\n\n Examples :\n listen http_proxy 0.0.0.0:80\n mode http\n option httplog\n option logasap\n log 192.168.2.200 local3\n\n >>> Feb 6 12:14:14 localhost \\n haproxy[14389]: 10.0.1.2:33317 [06/Feb/2009:12:14:14.655] http-in \\n static/srv1 9/10/7/14/+30 200 +243 - - ---- 3/1/1/1/0 1/0 \\n "GET /image.iso HTTP/1.0"\n\n See also : "option httplog", "capture response header", and section 8 about\n logging.\n\n\noption mysql-check [ user <username> ]\n Use MySQL health checks for server testing\n May be used in sections : defaults | frontend | listen | backend\n yes | no | yes | yes\n Arguments :\n <username> This is the username which will be used when connecting to MySQL\n server.\n\n If you specify a username, the check consists of sending two MySQL packet,\n one Client Authentication packet, and one QUIT packet, to correctly close\n MySQL session. We then parse the MySQL Handshake Initialisation packet and/or\n Error packet. It is a basic but useful test which does not produce error nor\n aborted connect on the server. However, it requires adding an authorization\n in the MySQL table, like this :\n\n USE mysql;\n INSERT INTO user (Host,User) values ('<ip_of_haproxy>','<username>');\n FLUSH PRIVILEGES;\n\n If you don't specify a username (it is deprecated and not recommended), the\n check only consists in parsing the Mysql Handshake Initialisation packet or\n Error packet, we don't send anything in this mode. It was reported that it\n can generate lockout if check is too frequent and/or if there is not enough\n traffic. In fact, you need in this case to check MySQL "max_connect_errors"\n value as if a connection is established successfully within fewer than MySQL\n "max_connect_errors" attempts after a previous connection was interrupted,\n the error count for the host is cleared to zero. If HAProxy's server get\n blocked, the "FLUSH HOSTS" statement is the only way to unblock it.\n\n Remember that this does not check database presence nor database consistency.\n To do this, you can use an external check with xinetd for example.\n\n The check requires MySQL >=3.22, for older version, please use TCP check.\n\n Most often, an incoming MySQL server needs to see the client's IP address for\n various purposes, including IP privilege matching and connection logging.\n When possible, it is often wise to masquerade the client's IP address when\n connecting to the server using the "usesrc" argument of the "source" keyword,\n which requires the cttproxy feature to be compiled in, and the MySQL server\n to route the client via the machine hosting haproxy.\n\n See also: "option httpchk"\n\n\noption nolinger\nno option nolinger\n Enable or disable immediate session resource cleaning after close\n May be used in sections: defaults | frontend | listen | backend\n yes | yes | yes | yes\n Arguments : none\n\n When clients or servers abort connections in a dirty way (eg: they are\n physically disconnected), the session timeouts triggers and the session is\n closed. But it will remain in FIN_WAIT1 state for some time in the system,\n using some resources and possibly limiting the ability to establish newer\n connections.\n\n When this happens, it is possible to activate "option nolinger" which forces\n the system to immediately remove any socket's pending data on close. Thus,\n the session is instantly purged from the system's tables. This usually has\n side effects such as increased number of TCP resets due to old retransmits\n getting immediately rejected. Some firewalls may sometimes complain about\n this too.\n\n For this reason, it is not recommended to use this option when not absolutely\n needed. You know that you need it when you have thousands of FIN_WAIT1\n sessions on your system (TIME_WAIT ones do not count).\n\n This option may be used both on frontends and backends, depending on the side\n where it is required. Use it on the frontend for clients, and on the backend\n for servers.\n\n If this option has been enabled in a "defaults" section, it can be disabled\n in a specific instance by prepending the "no" keyword before it.\n\n\noption originalto [ except <network> ] [ header <name> ]\n Enable insertion of the X-Original-To header to requests sent to servers\n May be used in sections : defaults | frontend | listen | backend\n yes | yes | yes | yes\n Arguments :\n <network> is an optional argument used to disable this option for sources\n matching <network>\n <name> an optional argument to specify a different "X-Original-To"\n header name.\n\n Since HAProxy can work in transparent mode, every request from a client can\n be redirected to the proxy and HAProxy itself can proxy every request to a\n complex SQUID environment and the destination host from SO_ORIGINAL_DST will\n be lost. This is annoying when you want access rules based on destination ip\n addresses. To solve this problem, a new HTTP header "X-Original-To" may be\n added by HAProxy to all requests sent to the server. This header contains a\n value representing the original destination IP address. Since this must be\n configured to always use the last occurrence of this header only. Note that\n only the last occurrence of the header must be used, since it is really\n possible that the client has already brought one.\n\n The keyword "header" may be used to supply a different header name to replace\n the default "X-Original-To". This can be useful where you might already\n have a "X-Original-To" header from a different application, and you need\n preserve it. Also if your backend server doesn't use the "X-Original-To"\n header and requires different one.\n\n Sometimes, a same HAProxy instance may be shared between a direct client\n access and a reverse-proxy access (for instance when an SSL reverse-proxy is\n used to decrypt HTTPS traffic). It is possible to disable the addition of the\n header for a known source address or network by adding the "except" keyword\n followed by the network address. In this case, any source IP matching the\n network will not cause an addition of this header. Most common uses are with\n private networks or 127.0.0.1.\n\n This option may be specified either in the frontend or in the backend. If at\n least one of them uses it, the header will be added. Note that the backend's\n setting of the header subargument takes precedence over the frontend's if\n both are defined.\n\n It is important to note that by default, HAProxy works in tunnel mode and\n only inspects the first request of a connection, meaning that only the first\n request will have the header appended, which is certainly not what you want.\n In order to fix this, ensure that any of the "httpclose", "forceclose" or\n "http-server-close" options is set when using this option.\n\n Examples :\n # Original Destination address\n frontend www\n mode http\n option originalto except 127.0.0.1\n\n # Those servers want the IP Address in X-Client-Dst\n backend www\n mode http\n option originalto header X-Client-Dst\n\n See also : "option httpclose", "option http-server-close",\n "option forceclose"\n\n\noption persist\nno option persist\n Enable or disable forced persistence on down servers\n May be used in sections: defaults | frontend | listen | backend\n yes | no | yes | yes\n Arguments : none\n\n When an HTTP request reaches a backend with a cookie which references a dead\n server, by default it is redispatched to another server. It is possible to\n force the request to be sent to the dead server first using "option persist"\n if absolutely needed. A common use case is when servers are under extreme\n load and spend their time flapping. In this case, the users would still be\n directed to the server they opened the session on, in the hope they would be\n correctly served. It is recommended to use "option redispatch" in conjunction\n with this option so that in the event it would not be possible to connect to\n the server at all (server definitely dead), the client would finally be\n redirected to another valid server.\n\n If this option has been enabled in a "defaults" section, it can be disabled\n in a specific instance by prepending the "no" keyword before it.\n\n See also : "option redispatch", "retries", "force-persist"\n\n\noption redispatch\nno option redispatch\n Enable or disable session redistribution in case of connection failure\n May be used in sections: defaults | frontend | listen | backend\n yes | no | yes | yes\n Arguments : none\n\n In HTTP mode, if a server designated by a cookie is down, clients may\n definitely stick to it because they cannot flush the cookie, so they will not\n be able to access the service anymore.\n\n Specifying "option redispatch" will allow the proxy to break their\n persistence and redistribute them to a working server.\n\n It also allows to retry last connection to another server in case of multiple\n connection failures. Of course, it requires having "retries" set to a nonzero\n value.\n\n This form is the preferred form, which replaces both the "redispatch" and\n "redisp" keywords.\n\n If this option has been enabled in a "defaults" section, it can be disabled\n in a specific instance by prepending the "no" keyword before it.\n\n See also : "redispatch", "retries", "force-persist"\n\n\noption smtpchk\noption smtpchk <hello> <domain>\n Use SMTP health checks for server testing\n May be used in sections : defaults | frontend | listen | backend\n yes | no | yes | yes\n Arguments :\n <hello> is an optional argument. It is the "hello" command to use. It can\n be either "HELO" (for SMTP) or "EHLO" (for ESTMP). All other\n values will be turned into the default command ("HELO").\n\n <domain> is the domain name to present to the server. It may only be\n specified (and is mandatory) if the hello command has been\n specified. By default, "localhost" is used.\n\n When "option smtpchk" is set, the health checks will consist in TCP\n connections followed by an SMTP command. By default, this command is\n "HELO localhost". The server's return code is analyzed and only return codes\n starting with a "2" will be considered as valid. All other responses,\n including a lack of response will constitute an error and will indicate a\n dead server.\n\n This test is meant to be used with SMTP servers or relays. Depending on the\n request, it is possible that some servers do not log each connection attempt,\n so you may want to experiment to improve the behaviour. Using telnet on port\n 25 is often easier than adjusting the configuration.\n\n Most often, an incoming SMTP server needs to see the client's IP address for\n various purposes, including spam filtering, anti-spoofing and logging. When\n possible, it is often wise to masquerade the client's IP address when\n connecting to the server using the "usesrc" argument of the "source" keyword,\n which requires the cttproxy feature to be compiled in.\n\n Example :\n option smtpchk HELO mydomain.org\n\n See also : "option httpchk", "source"\n\n\noption socket-stats\nno option socket-stats\n\n Enable or disable collecting & providing separate statistics for each socket.\n May be used in sections : defaults | frontend | listen | backend\n yes | yes | yes | no\n\n Arguments : none\n\n\noption splice-auto\nno option splice-auto\n Enable or disable automatic kernel acceleration on sockets in both directions\n May be used in sections : defaults | frontend | listen | backend\n yes | yes | yes | yes\n Arguments : none\n\n When this option is enabled either on a frontend or on a backend, haproxy\n will automatically evaluate the opportunity to use kernel tcp splicing to\n forward data between the client and the server, in either direction. Haproxy\n uses heuristics to estimate if kernel splicing might improve performance or\n not. Both directions are handled independently. Note that the heuristics used\n are not much aggressive in order to limit excessive use of splicing. This\n option requires splicing to be enabled at compile time, and may be globally\n disabled with the global option "nosplice". Since splice uses pipes, using it\n requires that there are enough spare pipes.\n\n Important note: kernel-based TCP splicing is a Linux-specific feature which\n first appeared in kernel 2.6.25. It offers kernel-based acceleration to\n transfer data between sockets without copying these data to user-space, thus\n providing noticeable performance gains and CPU cycles savings. Since many\n early implementations are buggy, corrupt data and/or are inefficient, this\n feature is not enabled by default, and it should be used with extreme care.\n While it is not possible to detect the correctness of an implementation,\n 2.6.29 is the first version offering a properly working implementation. In\n case of doubt, splicing may be globally disabled using the global "nosplice"\n keyword.\n\n Example :\n option splice-auto\n\n If this option has been enabled in a "defaults" section, it can be disabled\n in a specific instance by prepending the "no" keyword before it.\n\n See also : "option splice-request", "option splice-response", and global\n options "nosplice" and "maxpipes"\n\n\noption splice-request\nno option splice-request\n Enable or disable automatic kernel acceleration on sockets for requests\n May be used in sections : defaults | frontend | listen | backend\n yes | yes | yes | yes\n Arguments : none\n\n When this option is enabled either on a frontend or on a backend, haproxy\n will user kernel tcp splicing whenever possible to forward data going from\n the client to the server. It might still use the recv/send scheme if there\n are no spare pipes left. This option requires splicing to be enabled at\n compile time, and may be globally disabled with the global option "nosplice".\n Since splice uses pipes, using it requires that there are enough spare pipes.\n\n Important note: see "option splice-auto" for usage limitations.\n\n Example :\n option splice-request\n\n If this option has been enabled in a "defaults" section, it can be disabled\n in a specific instance by prepending the "no" keyword before it.\n\n See also : "option splice-auto", "option splice-response", and global options\n "nosplice" and "maxpipes"\n\n\noption splice-response\nno option splice-response\n Enable or disable automatic kernel acceleration on sockets for responses\n May be used in sections : defaults | frontend | listen | backend\n yes | yes | yes | yes\n Arguments : none\n\n When this option is enabled either on a frontend or on a backend, haproxy\n will user kernel tcp splicing whenever possible to forward data going from\n the server to the client. It might still use the recv/send scheme if there\n are no spare pipes left. This option requires splicing to be enabled at\n compile time, and may be globally disabled with the global option "nosplice".\n Since splice uses pipes, using it requires that there are enough spare pipes.\n\n Important note: see "option splice-auto" for usage limitations.\n\n Example :\n option splice-response\n\n If this option has been enabled in a "defaults" section, it can be disabled\n in a specific instance by prepending the "no" keyword before it.\n\n See also : "option splice-auto", "option splice-request", and global options\n "nosplice" and "maxpipes"\n\n\noption srvtcpka\nno option srvtcpka\n Enable or disable the sending of TCP keepalive packets on the server side\n May be used in sections : defaults | frontend | listen | backend\n yes | no | yes | yes\n Arguments : none\n\n When there is a firewall or any session-aware component between a client and\n a server, and when the protocol involves very long sessions with long idle\n periods (eg: remote desktops), there is a risk that one of the intermediate\n components decides to expire a session which has remained idle for too long.\n\n Enabling socket-level TCP keep-alives makes the system regularly send packets\n to the other end of the connection, leaving it active. The delay between\n keep-alive probes is controlled by the system only and depends both on the\n operating system and its tuning parameters.\n\n It is important to understand that keep-alive packets are neither emitted nor\n received at the application level. It is only the network stacks which sees\n them. For this reason, even if one side of the proxy already uses keep-alives\n to maintain its connection alive, those keep-alive packets will not be\n forwarded to the other side of the proxy.\n\n Please note that this has nothing to do with HTTP keep-alive.\n\n Using option "srvtcpka" enables the emission of TCP keep-alive probes on the\n server side of a connection, which should help when session expirations are\n noticed between HAProxy and a server.\n\n If this option has been enabled in a "defaults" section, it can be disabled\n in a specific instance by prepending the "no" keyword before it.\n\n See also : "option clitcpka", "option tcpka"\n\n\noption ssl-hello-chk\n Use SSLv3 client hello health checks for server testing\n May be used in sections : defaults | frontend | listen | backend\n yes | no | yes | yes\n Arguments : none\n\n When some SSL-based protocols are relayed in TCP mode through HAProxy, it is\n possible to test that the server correctly talks SSL instead of just testing\n that it accepts the TCP connection. When "option ssl-hello-chk" is set, pure\n SSLv3 client hello messages are sent once the connection is established to\n the server, and the response is analyzed to find an SSL server hello message.\n The server is considered valid only when the response contains this server\n hello message.\n\n All servers tested till there correctly reply to SSLv3 client hello messages,\n and most servers tested do not even log the requests containing only hello\n messages, which is appreciable.\n\n See also: "option httpchk"\n\n\noption tcp-smart-accept\nno option tcp-smart-accept\n Enable or disable the saving of one ACK packet during the accept sequence\n May be used in sections : defaults | frontend | listen | backend\n yes | yes | yes | no\n Arguments : none\n\n When an HTTP connection request comes in, the system acknowledges it on\n behalf of HAProxy, then the client immediately sends its request, and the\n system acknowledges it too while it is notifying HAProxy about the new\n connection. HAProxy then reads the request and responds. This means that we\n have one TCP ACK sent by the system for nothing, because the request could\n very well be acknowledged by HAProxy when it sends its response.\n\n For this reason, in HTTP mode, HAProxy automatically asks the system to avoid\n sending this useless ACK on platforms which support it (currently at least\n Linux). It must not cause any problem, because the system will send it anyway\n after 40 ms if the response takes more time than expected to come.\n\n During complex network debugging sessions, it may be desirable to disable\n this optimization because delayed ACKs can make troubleshooting more complex\n when trying to identify where packets are delayed. It is then possible to\n fall back to normal behaviour by specifying "no option tcp-smart-accept".\n\n It is also possible to force it for non-HTTP proxies by simply specifying\n "option tcp-smart-accept". For instance, it can make sense with some services\n such as SMTP where the server speaks first.\n\n It is recommended to avoid forcing this option in a defaults section. In case\n of doubt, consider setting it back to automatic values by prepending the\n "default" keyword before it, or disabling it using the "no" keyword.\n\n See also : "option tcp-smart-connect"\n\n\noption tcp-smart-connect\nno option tcp-smart-connect\n Enable or disable the saving of one ACK packet during the connect sequence\n May be used in sections : defaults | frontend | listen | backend\n yes | no | yes | yes\n Arguments : none\n\n On certain systems (at least Linux), HAProxy can ask the kernel not to\n immediately send an empty ACK upon a connection request, but to directly\n send the buffer request instead. This saves one packet on the network and\n thus boosts performance. It can also be useful for some servers, because they\n immediately get the request along with the incoming connection.\n\n This feature is enabled when "option tcp-smart-connect" is set in a backend.\n It is not enabled by default because it makes network troubleshooting more\n complex.\n\n It only makes sense to enable it with protocols where the client speaks first\n such as HTTP. In other situations, if there is no data to send in place of\n the ACK, a normal ACK is sent.\n\n If this option has been enabled in a "defaults" section, it can be disabled\n in a specific instance by prepending the "no" keyword before it.\n\n See also : "option tcp-smart-accept"\n\n\noption tcpka\n Enable or disable the sending of TCP keepalive packets on both sides\n May be used in sections : defaults | frontend | listen | backend\n yes | yes | yes | yes\n Arguments : none\n\n When there is a firewall or any session-aware component between a client and\n a server, and when the protocol involves very long sessions with long idle\n periods (eg: remote desktops), there is a risk that one of the intermediate\n components decides to expire a session which has remained idle for too long.\n\n Enabling socket-level TCP keep-alives makes the system regularly send packets\n to the other end of the connection, leaving it active. The delay between\n keep-alive probes is controlled by the system only and depends both on the\n operating system and its tuning parameters.\n\n It is important to understand that keep-alive packets are neither emitted nor\n received at the application level. It is only the network stacks which sees\n them. For this reason, even if one side of the proxy already uses keep-alives\n to maintain its connection alive, those keep-alive packets will not be\n forwarded to the other side of the proxy.\n\n Please note that this has nothing to do with HTTP keep-alive.\n\n Using option "tcpka" enables the emission of TCP keep-alive probes on both\n the client and server sides of a connection. Note that this is meaningful\n only in "defaults" or "listen" sections. If this option is used in a\n frontend, only the client side will get keep-alives, and if this option is\n used in a backend, only the server side will get keep-alives. For this\n reason, it is strongly recommended to explicitly use "option clitcpka" and\n "option srvtcpka" when the configuration is split between frontends and\n backends.\n\n See also : "option clitcpka", "option srvtcpka"\n\n\noption tcplog\n Enable advanced logging of TCP connections with session state and timers\n May be used in sections : defaults | frontend | listen | backend\n yes | yes | yes | yes\n Arguments : none\n\n By default, the log output format is very poor, as it only contains the\n source and destination addresses, and the instance name. By specifying\n "option tcplog", each log line turns into a much richer format including, but\n not limited to, the connection timers, the session status, the connections\n numbers, the frontend, backend and server name, and of course the source\n address and ports. This option is useful for pure TCP proxies in order to\n find which of the client or server disconnects or times out. For normal HTTP\n proxies, it's better to use "option httplog" which is even more complete.\n\n This option may be set either in the frontend or the backend.\n\n See also : "option httplog", and section 8 about logging.\n\n\noption transparent\nno option transparent\n Enable client-side transparent proxying\n May be used in sections : defaults | frontend | listen | backend\n yes | no | yes | yes\n Arguments : none\n\n This option was introduced in order to provide layer 7 persistence to layer 3\n load balancers. The idea is to use the OS's ability to redirect an incoming\n connection for a remote address to a local process (here HAProxy), and let\n this process know what address was initially requested. When this option is\n used, sessions without cookies will be forwarded to the original destination\n IP address of the incoming request (which should match that of another\n equipment), while requests with cookies will still be forwarded to the\n appropriate server.\n\n Note that contrary to a common belief, this option does NOT make HAProxy\n present the client's IP to the server when establishing the connection.\n\n See also: the "usesrc" argument of the "source" keyword, and the\n "transparent" option of the "bind" keyword.\n\n\npersist rdp-cookie\npersist rdp-cookie(name)\n Enable RDP cookie-based persistence\n May be used in sections : defaults | frontend | listen | backend\n yes | no | yes | yes\n Arguments :\n <name> is the optional name of the RDP cookie to check. If omitted, the\n default cookie name "msts" will be used. There currently is no\n valid reason to change this name.\n\n This statement enables persistence based on an RDP cookie. The RDP cookie\n contains all information required to find the server in the list of known\n servers. So when this option is set in the backend, the request is analysed\n and if an RDP cookie is found, it is decoded. If it matches a known server\n which is still UP (or if "option persist" is set), then the connection is\n forwarded to this server.\n\n Note that this only makes sense in a TCP backend, but for this to work, the\n frontend must have waited long enough to ensure that an RDP cookie is present\n in the request buffer. This is the same requirement as with the "rdp-cookie"\n load-balancing method. Thus it is highly recommended to put all statements in\n a single "listen" section.\n\n Also, it is important to understand that the terminal server will emit this\n RDP cookie only if it is configured for "token redirection mode", which means\n that the "IP address redirection" option is disabled.\n\n Example :\n listen tse-farm\n bind :3389\n # wait up to 5s for an RDP cookie in the request\n tcp-request inspect-delay 5s\n tcp-request content accept if RDP_COOKIE\n # apply RDP cookie persistence\n persist rdp-cookie\n # if server is unknown, let's balance on the same cookie.\n # alternatively, "balance leastconn" may be useful too.\n balance rdp-cookie\n server srv1 1.1.1.1:3389\n server srv2 1.1.1.2:3389\n\n See also : "balance rdp-cookie", "tcp-request" and the "req_rdp_cookie" ACL.\n\n\nrate-limit sessions <rate>\n Set a limit on the number of new sessions accepted per second on a frontend\n May be used in sections : defaults | frontend | listen | backend\n yes | yes | yes | no\n Arguments :\n <rate> The <rate> parameter is an integer designating the maximum number\n of new sessions per second to accept on the frontend.\n\n When the frontend reaches the specified number of new sessions per second, it\n stops accepting new connections until the rate drops below the limit again.\n During this time, the pending sessions will be kept in the socket's backlog\n (in system buffers) and haproxy will not even be aware that sessions are\n pending. When applying very low limit on a highly loaded service, it may make\n sense to increase the socket's backlog using the "backlog" keyword.\n\n This feature is particularly efficient at blocking connection-based attacks\n or service abuse on fragile servers. Since the session rate is measured every\n millisecond, it is extremely accurate. Also, the limit applies immediately,\n no delay is needed at all to detect the threshold.\n\n Example : limit the connection rate on SMTP to 10 per second max\n listen smtp\n mode tcp\n bind :25\n rate-limit sessions 10\n server 127.0.0.1:1025\n\n Note : when the maximum rate is reached, the frontend's status appears as\n "FULL" in the statistics, exactly as when it is saturated.\n\n See also : the "backlog" keyword and the "fe_sess_rate" ACL criterion.\n\n\nredirect location <to> [code <code>] <option> [{if | unless} <condition>]\nredirect prefix <to> [code <code>] <option> [{if | unless} <condition>]\n Return an HTTP redirection if/unless a condition is matched\n May be used in sections : defaults | frontend | listen | backend\n no | yes | yes | yes\n\n If/unless the condition is matched, the HTTP request will lead to a redirect\n response. If no condition is specified, the redirect applies unconditionally.\n\n Arguments :\n <to> With "redirect location", the exact value in <to> is placed into\n the HTTP "Location" header. In case of "redirect prefix", the\n "Location" header is built from the concatenation of <to> and the\n complete URI, including the query string, unless the "drop-query"\n option is specified (see below). As a special case, if <to>\n equals exactly "/" in prefix mode, then nothing is inserted\n before the original URI. It allows one to redirect to the same\n URL.\n\n <code> The code is optional. It indicates which type of HTTP redirection\n is desired. Only codes 301, 302 and 303 are supported, and 302 is\n used if no code is specified. 301 means "Moved permanently", and\n a browser may cache the Location. 302 means "Moved permanently"\n and means that the browser should not cache the redirection. 303\n is equivalent to 302 except that the browser will fetch the\n location with a GET method.\n\n <option> There are several options which can be specified to adjust the\n expected behaviour of a redirection :\n\n - "drop-query"\n When this keyword is used in a prefix-based redirection, then the\n location will be set without any possible query-string, which is useful\n for directing users to a non-secure page for instance. It has no effect\n with a location-type redirect.\n\n - "append-slash"\n This keyword may be used in conjunction with "drop-query" to redirect\n users who use a URL not ending with a '/' to the same one with the '/'.\n It can be useful to ensure that search engines will only see one URL.\n For this, a return code 301 is preferred.\n\n - "set-cookie NAME[=value]"\n A "Set-Cookie" header will be added with NAME (and optionally "=value")\n to the response. This is sometimes used to indicate that a user has\n been seen, for instance to protect against some types of DoS. No other\n cookie option is added, so the cookie will be a session cookie. Note\n that for a browser, a sole cookie name without an equal sign is\n different from a cookie with an equal sign.\n\n - "clear-cookie NAME[=]"\n A "Set-Cookie" header will be added with NAME (and optionally "="), but\n with the "Max-Age" attribute set to zero. This will tell the browser to\n delete this cookie. It is useful for instance on logout pages. It is\n important to note that clearing the cookie "NAME" will not remove a\n cookie set with "NAME=value". You have to clear the cookie "NAME=" for\n that, because the browser makes the difference.\n\n Example: move the login URL only to HTTPS.\n acl clear dst_port 80\n acl secure dst_port 8080\n acl login_page url_beg /login\n acl logout url_beg /logout\n acl uid_given url_reg /login?userid=[^&]+\n acl cookie_set hdr_sub(cookie) SEEN=1\n\n redirect prefix https://mysite.com set-cookie SEEN=1 if !cookie_set\n redirect prefix https://mysite.com if login_page !secure\n redirect prefix http://mysite.com drop-query if login_page !uid_given\n redirect location http://mysite.com/ if !login_page secure\n redirect location / clear-cookie USERID= if logout\n\n Example: send redirects for request for articles without a '/'.\n acl missing_slash path_reg ^/article/[^/]*$\n redirect code 301 prefix / drop-query append-slash if missing_slash\n\n See section 7 about ACL usage.\n\n\nredisp (deprecated)\nredispatch (deprecated)\n Enable or disable session redistribution in case of connection failure\n May be used in sections: defaults | frontend | listen | backend\n yes | no | yes | yes\n Arguments : none\n\n In HTTP mode, if a server designated by a cookie is down, clients may\n definitely stick to it because they cannot flush the cookie, so they will not\n be able to access the service anymore.\n\n Specifying "redispatch" will allow the proxy to break their persistence and\n redistribute them to a working server.\n\n It also allows to retry last connection to another server in case of multiple\n connection failures. Of course, it requires having "retries" set to a nonzero\n value.\n\n This form is deprecated, do not use it in any new configuration, use the new\n "option redispatch" instead.\n\n See also : "option redispatch"\n\n\nreqadd <string> [{if | unless} <cond>]\n Add a header at the end of the HTTP request\n May be used in sections : defaults | frontend | listen | backend\n no | yes | yes | yes\n Arguments :\n <string> is the complete line to be added. Any space or known delimiter\n must be escaped using a backslash ('\'). Please refer to section\n 6 about HTTP header manipulation for more information.\n\n <cond> is an optional matching condition built from ACLs. It makes it\n possible to ignore this rule when other conditions are not met.\n\n A new line consisting in <string> followed by a line feed will be added after\n the last header of an HTTP request.\n\n Header transformations only apply to traffic which passes through HAProxy,\n and not to traffic generated by HAProxy, such as health-checks or error\n responses.\n\n Example : add "X-Proto: SSL" to requests coming via port 81\n acl is-ssl dst_port 81\n reqadd X-Proto:\ SSL if is-ssl\n\n See also: "rspadd", section 6 about HTTP header manipulation, and section 7\n about ACLs.\n\n\nreqallow <search> [{if | unless} <cond>]\nreqiallow <search> [{if | unless} <cond>] (ignore case)\n Definitely allow an HTTP request if a line matches a regular expression\n May be used in sections : defaults | frontend | listen | backend\n no | yes | yes | yes\n Arguments :\n <search> is the regular expression applied to HTTP headers and to the\n request line. This is an extended regular expression. Parenthesis\n grouping is supported and no preliminary backslash is required.\n Any space or known delimiter must be escaped using a backslash\n ('\'). The pattern applies to a full line at a time. The\n "reqallow" keyword strictly matches case while "reqiallow"\n ignores case.\n\n <cond> is an optional matching condition built from ACLs. It makes it\n possible to ignore this rule when other conditions are not met.\n\n A request containing any line which matches extended regular expression\n <search> will mark the request as allowed, even if any later test would\n result in a deny. The test applies both to the request line and to request\n headers. Keep in mind that URLs in request line are case-sensitive while\n header names are not.\n\n It is easier, faster and more powerful to use ACLs to write access policies.\n Reqdeny, reqallow and reqpass should be avoided in new designs.\n\n Example :\n # allow www.* but refuse *.local\n reqiallow ^Host:\ www\.\n reqideny ^Host:\ .*\.local\n\n See also: "reqdeny", "block", section 6 about HTTP header manipulation, and\n section 7 about ACLs.\n\n\nreqdel <search> [{if | unless} <cond>]\nreqidel <search> [{if | unless} <cond>] (ignore case)\n Delete all headers matching a regular expression in an HTTP request\n May be used in sections : defaults | frontend | listen | backend\n no | yes | yes | yes\n Arguments :\n <search> is the regular expression applied to HTTP headers and to the\n request line. This is an extended regular expression. Parenthesis\n grouping is supported and no preliminary backslash is required.\n Any space or known delimiter must be escaped using a backslash\n ('\'). The pattern applies to a full line at a time. The "reqdel"\n keyword strictly matches case while "reqidel" ignores case.\n\n <cond> is an optional matching condition built from ACLs. It makes it\n possible to ignore this rule when other conditions are not met.\n\n Any header line matching extended regular expression <search> in the request\n will be completely deleted. Most common use of this is to remove unwanted\n and/or dangerous headers or cookies from a request before passing it to the\n next servers.\n\n Header transformations only apply to traffic which passes through HAProxy,\n and not to traffic generated by HAProxy, such as health-checks or error\n responses. Keep in mind that header names are not case-sensitive.\n\n Example :\n # remove X-Forwarded-For header and SERVER cookie\n reqidel ^X-Forwarded-For:.*\n reqidel ^Cookie:.*SERVER=\n\n See also: "reqadd", "reqrep", "rspdel", section 6 about HTTP header\n manipulation, and section 7 about ACLs.\n\n\nreqdeny <search> [{if | unless} <cond>]\nreqideny <search> [{if | unless} <cond>] (ignore case)\n Deny an HTTP request if a line matches a regular expression\n May be used in sections : defaults | frontend | listen | backend\n no | yes | yes | yes\n Arguments :\n <search> is the regular expression applied to HTTP headers and to the\n request line. This is an extended regular expression. Parenthesis\n grouping is supported and no preliminary backslash is required.\n Any space or known delimiter must be escaped using a backslash\n ('\'). The pattern applies to a full line at a time. The\n "reqdeny" keyword strictly matches case while "reqideny" ignores\n case.\n\n <cond> is an optional matching condition built from ACLs. It makes it\n possible to ignore this rule when other conditions are not met.\n\n A request containing any line which matches extended regular expression\n <search> will mark the request as denied, even if any later test would\n result in an allow. The test applies both to the request line and to request\n headers. Keep in mind that URLs in request line are case-sensitive while\n header names are not.\n\n A denied request will generate an "HTTP 403 forbidden" response once the\n complete request has been parsed. This is consistent with what is practiced\n using ACLs.\n\n It is easier, faster and more powerful to use ACLs to write access policies.\n Reqdeny, reqallow and reqpass should be avoided in new designs.\n\n Example :\n # refuse *.local, then allow www.*\n reqideny ^Host:\ .*\.local\n reqiallow ^Host:\ www\.\n\n See also: "reqallow", "rspdeny", "block", section 6 about HTTP header\n manipulation, and section 7 about ACLs.\n\n\nreqpass <search> [{if | unless} <cond>]\nreqipass <search> [{if | unless} <cond>] (ignore case)\n Ignore any HTTP request line matching a regular expression in next rules\n May be used in sections : defaults | frontend | listen | backend\n no | yes | yes | yes\n Arguments :\n <search> is the regular expression applied to HTTP headers and to the\n request line. This is an extended regular expression. Parenthesis\n grouping is supported and no preliminary backslash is required.\n Any space or known delimiter must be escaped using a backslash\n ('\'). The pattern applies to a full line at a time. The\n "reqpass" keyword strictly matches case while "reqipass" ignores\n case.\n\n <cond> is an optional matching condition built from ACLs. It makes it\n possible to ignore this rule when other conditions are not met.\n\n A request containing any line which matches extended regular expression\n <search> will skip next rules, without assigning any deny or allow verdict.\n The test applies both to the request line and to request headers. Keep in\n mind that URLs in request line are case-sensitive while header names are not.\n\n It is easier, faster and more powerful to use ACLs to write access policies.\n Reqdeny, reqallow and reqpass should be avoided in new designs.\n\n Example :\n # refuse *.local, then allow www.*, but ignore "www.private.local"\n reqipass ^Host:\ www.private\.local\n reqideny ^Host:\ .*\.local\n reqiallow ^Host:\ www\.\n\n See also: "reqallow", "reqdeny", "block", section 6 about HTTP header\n manipulation, and section 7 about ACLs.\n\n\nreqrep <search> <string> [{if | unless} <cond>]\nreqirep <search> <string> [{if | unless} <cond>] (ignore case)\n Replace a regular expression with a string in an HTTP request line\n May be used in sections : defaults | frontend | listen | backend\n no | yes | yes | yes\n Arguments :\n <search> is the regular expression applied to HTTP headers and to the\n request line. This is an extended regular expression. Parenthesis\n grouping is supported and no preliminary backslash is required.\n Any space or known delimiter must be escaped using a backslash\n ('\'). The pattern applies to a full line at a time. The "reqrep"\n keyword strictly matches case while "reqirep" ignores case.\n\n <string> is the complete line to be added. Any space or known delimiter\n must be escaped using a backslash ('\'). References to matched\n pattern groups are possible using the common \N form, with N\n being a single digit between 0 and 9. Please refer to section\n 6 about HTTP header manipulation for more information.\n\n <cond> is an optional matching condition built from ACLs. It makes it\n possible to ignore this rule when other conditions are not met.\n\n Any line matching extended regular expression <search> in the request (both\n the request line and header lines) will be completely replaced with <string>.\n Most common use of this is to rewrite URLs or domain names in "Host" headers.\n\n Header transformations only apply to traffic which passes through HAProxy,\n and not to traffic generated by HAProxy, such as health-checks or error\n responses. Note that for increased readability, it is suggested to add enough\n spaces between the request and the response. Keep in mind that URLs in\n request line are case-sensitive while header names are not.\n\n Example :\n # replace "/static/" with "/" at the beginning of any request path.\n reqrep ^([^\ :]*)\ /static/(.*) \1\ /\2\n # replace "www.mydomain.com" with "www" in the host name.\n reqirep ^Host:\ www.mydomain.com Host:\ www\n\n See also: "reqadd", "reqdel", "rsprep", section 6 about HTTP header\n manipulation, and section 7 about ACLs.\n\n\nreqtarpit <search> [{if | unless} <cond>]\nreqitarpit <search> [{if | unless} <cond>] (ignore case)\n Tarpit an HTTP request containing a line matching a regular expression\n May be used in sections : defaults | frontend | listen | backend\n no | yes | yes | yes\n Arguments :\n <search> is the regular expression applied to HTTP headers and to the\n request line. This is an extended regular expression. Parenthesis\n grouping is supported and no preliminary backslash is required.\n Any space or known delimiter must be escaped using a backslash\n ('\'). The pattern applies to a full line at a time. The\n "reqtarpit" keyword strictly matches case while "reqitarpit"\n ignores case.\n\n <cond> is an optional matching condition built from ACLs. It makes it\n possible to ignore this rule when other conditions are not met.\n\n A request containing any line which matches extended regular expression\n <search> will be tarpitted, which means that it will connect to nowhere, will\n be kept open for a pre-defined time, then will return an HTTP error 500 so\n that the attacker does not suspect it has been tarpitted. The status 500 will\n be reported in the logs, but the completion flags will indicate "PT". The\n delay is defined by "timeout tarpit", or "timeout connect" if the former is\n not set.\n\n The goal of the tarpit is to slow down robots attacking servers with\n identifiable requests. Many robots limit their outgoing number of connections\n and stay connected waiting for a reply which can take several minutes to\n come. Depending on the environment and attack, it may be particularly\n efficient at reducing the load on the network and firewalls.\n\n Examples :\n # ignore user-agents reporting any flavour of "Mozilla" or "MSIE", but\n # block all others.\n reqipass ^User-Agent:\.*(Mozilla|MSIE)\n reqitarpit ^User-Agent:\n\n # block bad guys\n acl badguys src 10.1.0.3 172.16.13.20/28\n reqitarpit . if badguys\n\n See also: "reqallow", "reqdeny", "reqpass", section 6 about HTTP header\n manipulation, and section 7 about ACLs.\n\n\nretries <value>\n Set the number of retries to perform on a server after a connection failure\n May be used in sections: defaults | frontend | listen | backend\n yes | no | yes | yes\n Arguments :\n <value> is the number of times a connection attempt should be retried on\n a server when a connection either is refused or times out. The\n default value is 3.\n\n It is important to understand that this value applies to the number of\n connection attempts, not full requests. When a connection has effectively\n been established to a server, there will be no more retry.\n\n In order to avoid immediate reconnections to a server which is restarting,\n a turn-around timer of 1 second is applied before a retry occurs.\n\n When "option redispatch" is set, the last retry may be performed on another\n server even if a cookie references a different server.\n\n See also : "option redispatch"\n\n\nrspadd <string> [{if | unless} <cond>]\n Add a header at the end of the HTTP response\n May be used in sections : defaults | frontend | listen | backend\n no | yes | yes | yes\n Arguments :\n <string> is the complete line to be added. Any space or known delimiter\n must be escaped using a backslash ('\'). Please refer to section\n 6 about HTTP header manipulation for more information.\n\n <cond> is an optional matching condition built from ACLs. It makes it\n possible to ignore this rule when other conditions are not met.\n\n A new line consisting in <string> followed by a line feed will be added after\n the last header of an HTTP response.\n\n Header transformations only apply to traffic which passes through HAProxy,\n and not to traffic generated by HAProxy, such as health-checks or error\n responses.\n\n See also: "reqadd", section 6 about HTTP header manipulation, and section 7\n about ACLs.\n\n\nrspdel <search> [{if | unless} <cond>]\nrspidel <search> [{if | unless} <cond>] (ignore case)\n Delete all headers matching a regular expression in an HTTP response\n May be used in sections : defaults | frontend | listen | backend\n no | yes | yes | yes\n Arguments :\n <search> is the regular expression applied to HTTP headers and to the\n response line. This is an extended regular expression, so\n parenthesis grouping is supported and no preliminary backslash\n is required. Any space or known delimiter must be escaped using\n a backslash ('\'). The pattern applies to a full line at a time.\n The "rspdel" keyword strictly matches case while "rspidel"\n ignores case.\n\n <cond> is an optional matching condition built from ACLs. It makes it\n possible to ignore this rule when other conditions are not met.\n\n Any header line matching extended regular expression <search> in the response\n will be completely deleted. Most common use of this is to remove unwanted\n and/or sensitive headers or cookies from a response before passing it to the\n client.\n\n Header transformations only apply to traffic which passes through HAProxy,\n and not to traffic generated by HAProxy, such as health-checks or error\n responses. Keep in mind that header names are not case-sensitive.\n\n Example :\n # remove the Server header from responses\n reqidel ^Server:.*\n\n See also: "rspadd", "rsprep", "reqdel", section 6 about HTTP header\n manipulation, and section 7 about ACLs.\n\n\nrspdeny <search> [{if | unless} <cond>]\nrspideny <search> [{if | unless} <cond>] (ignore case)\n Block an HTTP response if a line matches a regular expression\n May be used in sections : defaults | frontend | listen | backend\n no | yes | yes | yes\n Arguments :\n <search> is the regular expression applied to HTTP headers and to the\n response line. This is an extended regular expression, so\n parenthesis grouping is supported and no preliminary backslash\n is required. Any space or known delimiter must be escaped using\n a backslash ('\'). The pattern applies to a full line at a time.\n The "rspdeny" keyword strictly matches case while "rspideny"\n ignores case.\n\n <cond> is an optional matching condition built from ACLs. It makes it\n possible to ignore this rule when other conditions are not met.\n\n A response containing any line which matches extended regular expression\n <search> will mark the request as denied. The test applies both to the\n response line and to response headers. Keep in mind that header names are not\n case-sensitive.\n\n Main use of this keyword is to prevent sensitive information leak and to\n block the response before it reaches the client. If a response is denied, it\n will be replaced with an HTTP 502 error so that the client never retrieves\n any sensitive data.\n\n It is easier, faster and more powerful to use ACLs to write access policies.\n Rspdeny should be avoided in new designs.\n\n Example :\n # Ensure that no content type matching ms-word will leak\n rspideny ^Content-type:\.*/ms-word\n\n See also: "reqdeny", "acl", "block", section 6 about HTTP header manipulation\n and section 7 about ACLs.\n\n\nrsprep <search> <string> [{if | unless} <cond>]\nrspirep <search> <string> [{if | unless} <cond>] (ignore case)\n Replace a regular expression with a string in an HTTP response line\n May be used in sections : defaults | frontend | listen | backend\n no | yes | yes | yes\n Arguments :\n <search> is the regular expression applied to HTTP headers and to the\n response line. This is an extended regular expression, so\n parenthesis grouping is supported and no preliminary backslash\n is required. Any space or known delimiter must be escaped using\n a backslash ('\'). The pattern applies to a full line at a time.\n The "rsprep" keyword strictly matches case while "rspirep"\n ignores case.\n\n <string> is the complete line to be added. Any space or known delimiter\n must be escaped using a backslash ('\'). References to matched\n pattern groups are possible using the common \N form, with N\n being a single digit between 0 and 9. Please refer to section\n 6 about HTTP header manipulation for more information.\n\n <cond> is an optional matching condition built from ACLs. It makes it\n possible to ignore this rule when other conditions are not met.\n\n Any line matching extended regular expression <search> in the response (both\n the response line and header lines) will be completely replaced with\n <string>. Most common use of this is to rewrite Location headers.\n\n Header transformations only apply to traffic which passes through HAProxy,\n and not to traffic generated by HAProxy, such as health-checks or error\n responses. Note that for increased readability, it is suggested to add enough\n spaces between the request and the response. Keep in mind that header names\n are not case-sensitive.\n\n Example :\n # replace "Location: 127.0.0.1:8080" with "Location: www.mydomain.com"\n rspirep ^Location:\ 127.0.0.1:8080 Location:\ www.mydomain.com\n\n See also: "rspadd", "rspdel", "reqrep", section 6 about HTTP header\n manipulation, and section 7 about ACLs.\n\n\nserver <name> <address>[:port] [param*]\n Declare a server in a backend\n May be used in sections : defaults | frontend | listen | backend\n no | no | yes | yes\n Arguments :\n <name> is the internal name assigned to this server. This name will\n appear in logs and alerts. If "http-send-server-name" is\n set, it will be added to the request header sent to the server.\n\n <address> is the IPv4 address of the server. Alternatively, a resolvable\n hostname is supported, but this name will be resolved during\n start-up. If no address is specified in front of the port, or if\n address "0.0.0.0" is specified, then the address used will be the\n same as the one used by the incoming connection. This makes it\n possible to relay connections for many IPs on a different port or\n to perform transparent forwarding with connection limitation.\n\n <ports> is an optional port specification. If set, all connections will\n be sent to this port. If unset, the same port the client\n connected to will be used. The port may also be prefixed by a "+"\n or a "-". In this case, the server's port will be determined by\n adding this value to the client's port.\n\n <param*> is a list of parameters for this server. The "server" keywords\n accepts an important number of options and has a complete section\n dedicated to it. Please refer to section 5 for more details.\n\n Examples :\n server first 10.1.1.1:1080 cookie first check inter 1000\n server second 10.1.1.2:1080 cookie second check inter 1000\n\n See also: "default-server", "http-send-name-header" and section 5 about\n server options\n\n\nsource <addr>[:<port>] [usesrc { <addr2>[:<port2>] | client | clientip } ]\nsource <addr>[:<port>] [usesrc { <addr2>[:<port2>] | hdr_ip(<hdr>[,<occ>]) } ]\nsource <addr>[:<port>] [interface <name>]\n Set the source address for outgoing connections\n May be used in sections : defaults | frontend | listen | backend\n yes | no | yes | yes\n Arguments :\n <addr> is the IPv4 address HAProxy will bind to before connecting to a\n server. This address is also used as a source for health checks.\n The default value of 0.0.0.0 means that the system will select\n the most appropriate address to reach its destination.\n\n <port> is an optional port. It is normally not needed but may be useful\n in some very specific contexts. The default value of zero means\n the system will select a free port. Note that port ranges are not\n supported in the backend. If you want to force port ranges, you\n have to specify them on each "server" line.\n\n <addr2> is the IP address to present to the server when connections are\n forwarded in full transparent proxy mode. This is currently only\n supported on some patched Linux kernels. When this address is\n specified, clients connecting to the server will be presented\n with this address, while health checks will still use the address\n <addr>.\n\n <port2> is the optional port to present to the server when connections\n are forwarded in full transparent proxy mode (see <addr2> above).\n The default value of zero means the system will select a free\n port.\n\n <hdr> is the name of a HTTP header in which to fetch the IP to bind to.\n This is the name of a comma-separated header list which can\n contain multiple IP addresses. By default, the last occurrence is\n used. This is designed to work with the X-Forwarded-For header\n and to automatically bind to the the client's IP address as seen\n by previous proxy, typically Stunnel. In order to use another\n occurrence from the last one, please see the <occ> parameter\n below. When the header (or occurrence) is not found, no binding\n is performed so that the proxy's default IP address is used. Also\n keep in mind that the header name is case insensitive, as for any\n HTTP header.\n\n <occ> is the occurrence number of a value to be used in a multi-value\n header. This is to be used in conjunction with "hdr_ip(<hdr>)",\n in order to specificy which occurrence to use for the source IP\n address. Positive values indicate a position from the first\n occurrence, 1 being the first one. Negative values indicate\n positions relative to the last one, -1 being the last one. This\n is helpful for situations where an X-Forwarded-For header is set\n at the entry point of an infrastructure and must be used several\n proxy layers away. When this value is not specified, -1 is\n assumed. Passing a zero here disables the feature.\n\n <name> is an optional interface name to which to bind to for outgoing\n traffic. On systems supporting this features (currently, only\n Linux), this allows one to bind all traffic to the server to\n this interface even if it is not the one the system would select\n based on routing tables. This should be used with extreme care.\n Note that using this option requires root privileges.\n\n The "source" keyword is useful in complex environments where a specific\n address only is allowed to connect to the servers. It may be needed when a\n private address must be used through a public gateway for instance, and it is\n known that the system cannot determine the adequate source address by itself.\n\n An extension which is available on certain patched Linux kernels may be used\n through the "usesrc" optional keyword. It makes it possible to connect to the\n servers with an IP address which does not belong to the system itself. This\n is called "full transparent proxy mode". For this to work, the destination\n servers have to route their traffic back to this address through the machine\n running HAProxy, and IP forwarding must generally be enabled on this machine.\n\n In this "full transparent proxy" mode, it is possible to force a specific IP\n address to be presented to the servers. This is not much used in fact. A more\n common use is to tell HAProxy to present the client's IP address. For this,\n there are two methods :\n\n - present the client's IP and port addresses. This is the most transparent\n mode, but it can cause problems when IP connection tracking is enabled on\n the machine, because a same connection may be seen twice with different\n states. However, this solution presents the huge advantage of not\n limiting the system to the 64k outgoing address+port couples, because all\n of the client ranges may be used.\n\n - present only the client's IP address and select a spare port. This\n solution is still quite elegant but slightly less transparent (downstream\n firewalls logs will not match upstream's). It also presents the downside\n of limiting the number of concurrent connections to the usual 64k ports.\n However, since the upstream and downstream ports are different, local IP\n connection tracking on the machine will not be upset by the reuse of the\n same session.\n\n Note that depending on the transparent proxy technology used, it may be\n required to force the source address. In fact, cttproxy version 2 requires an\n IP address in <addr> above, and does not support setting of "0.0.0.0" as the\n IP address because it creates NAT entries which much match the exact outgoing\n address. Tproxy version 4 and some other kernel patches which work in pure\n forwarding mode generally will not have this limitation.\n\n This option sets the default source for all servers in the backend. It may\n also be specified in a "defaults" section. Finer source address specification\n is possible at the server level using the "source" server option. Refer to\n section 5 for more information.\n\n Examples :\n backend private\n # Connect to the servers using our 192.168.1.200 source address\n source 192.168.1.200\n\n backend transparent_ssl1\n # Connect to the SSL farm from the client's source address\n source 192.168.1.200 usesrc clientip\n\n backend transparent_ssl2\n # Connect to the SSL farm from the client's source address and port\n # not recommended if IP conntrack is present on the local machine.\n source 192.168.1.200 usesrc client\n\n backend transparent_ssl3\n # Connect to the SSL farm from the client's source address. It\n # is more conntrack-friendly.\n source 192.168.1.200 usesrc clientip\n\n backend transparent_smtp\n # Connect to the SMTP farm from the client's source address/port\n # with Tproxy version 4.\n source 0.0.0.0 usesrc clientip\n\n backend transparent_http\n # Connect to the servers using the client's IP as seen by previous\n # proxy.\n source 0.0.0.0 usesrc hdr_ip(x-forwarded-for,-1)\n\n See also : the "source" server option in section 5, the Tproxy patches for\n the Linux kernel on www.balabit.com, the "bind" keyword.\n\n\nsrvtimeout <timeout> (deprecated)\n Set the maximum inactivity time on the server side.\n May be used in sections : defaults | frontend | listen | backend\n yes | no | yes | yes\n Arguments :\n <timeout> is the timeout value specified in milliseconds by default, but\n can be in any other unit if the number is suffixed by the unit,\n as explained at the top of this document.\n\n The inactivity timeout applies when the server is expected to acknowledge or\n send data. In HTTP mode, this timeout is particularly important to consider\n during the first phase of the server's response, when it has to send the\n headers, as it directly represents the server's processing time for the\n request. To find out what value to put there, it's often good to start with\n what would be considered as unacceptable response times, then check the logs\n to observe the response time distribution, and adjust the value accordingly.\n\n The value is specified in milliseconds by default, but can be in any other\n unit if the number is suffixed by the unit, as specified at the top of this\n document. In TCP mode (and to a lesser extent, in HTTP mode), it is highly\n recommended that the client timeout remains equal to the server timeout in\n order to avoid complex situations to debug. Whatever the expected server\n response times, it is a good practice to cover at least one or several TCP\n packet losses by specifying timeouts that are slightly above multiples of 3\n seconds (eg: 4 or 5 seconds minimum).\n\n This parameter is specific to backends, but can be specified once for all in\n "defaults" sections. This is in fact one of the easiest solutions not to\n forget about it. An unspecified timeout results in an infinite timeout, which\n is not recommended. Such a usage is accepted and works but reports a warning\n during startup because it may results in accumulation of expired sessions in\n the system if the system's timeouts are not configured either.\n\n This parameter is provided for compatibility but is currently deprecated.\n Please use "timeout server" instead.\n\n See also : "timeout server", "timeout client" and "clitimeout".\n\n\nstats admin { if | unless } <cond>\n Enable statistics admin level if/unless a condition is matched\n May be used in sections : defaults | frontend | listen | backend\n no | no | yes | yes\n\n This statement enables the statistics admin level if/unless a condition is\n matched.\n\n The admin level allows to enable/disable servers from the web interface. By\n default, statistics page is read-only for security reasons.\n\n Note : Consider not using this feature in multi-process mode (nbproc > 1)\n unless you know what you do : memory is not shared between the\n processes, which can result in random behaviours.\n\n Currently, the POST request is limited to the buffer size minus the reserved\n buffer space, which means that if the list of servers is too long, the\n request won't be processed. It is recommended to alter few servers at a\n time.\n\n Example :\n # statistics admin level only for localhost\n backend stats_localhost\n stats enable\n stats admin if LOCALHOST\n\n Example :\n # statistics admin level always enabled because of the authentication\n backend stats_auth\n stats enable\n stats auth admin:AdMiN123\n stats admin if TRUE\n\n Example :\n # statistics admin level depends on the authenticated user\n userlist stats-auth\n group admin users admin\n user admin insecure-password AdMiN123\n group readonly users haproxy\n user haproxy insecure-password haproxy\n\n backend stats_auth\n stats enable\n acl AUTH http_auth(stats-auth)\n acl AUTH_ADMIN http_auth_group(stats-auth) admin\n stats http-request auth unless AUTH\n stats admin if AUTH_ADMIN\n\n See also : "stats enable", "stats auth", "stats http-request", "nbproc",\n "bind-process", section 3.4 about userlists and section 7 about\n ACL usage.\n\n\nstats auth <user>:<passwd>\n Enable statistics with authentication and grant access to an account\n May be used in sections : defaults | frontend | listen | backend\n yes | no | yes | yes\n Arguments :\n <user> is a user name to grant access to\n\n <passwd> is the cleartext password associated to this user\n\n This statement enables statistics with default settings, and restricts access\n to declared users only. It may be repeated as many times as necessary to\n allow as many users as desired. When a user tries to access the statistics\n without a valid account, a "401 Forbidden" response will be returned so that\n the browser asks the user to provide a valid user and password. The real\n which will be returned to the browser is configurable using "stats realm".\n\n Since the authentication method is HTTP Basic Authentication, the passwords\n circulate in cleartext on the network. Thus, it was decided that the\n configuration file would also use cleartext passwords to remind the users\n that those ones should not be sensitive and not shared with any other account.\n\n It is also possible to reduce the scope of the proxies which appear in the\n report using "stats scope".\n\n Though this statement alone is enough to enable statistics reporting, it is\n recommended to set all other settings in order to avoid relying on default\n unobvious parameters.\n\n Example :\n # public access (limited to this backend only)\n backend public_www\n server srv1 192.168.0.1:80\n stats enable\n stats hide-version\n stats scope .\n stats uri /admin?stats\n stats realm Haproxy\ Statistics\n stats auth admin1:AdMiN123\n stats auth admin2:AdMiN321\n\n # internal monitoring access (unlimited)\n backend private_monitoring\n stats enable\n stats uri /admin?stats\n stats refresh 5s\n\n See also : "stats enable", "stats realm", "stats scope", "stats uri"\n\n\nstats enable\n Enable statistics reporting with default settings\n May be used in sections : defaults | frontend | listen | backend\n yes | no | yes | yes\n Arguments : none\n\n This statement enables statistics reporting with default settings defined\n at build time. Unless stated otherwise, these settings are used :\n - stats uri : /haproxy?stats\n - stats realm : "HAProxy Statistics"\n - stats auth : no authentication\n - stats scope : no restriction\n\n Though this statement alone is enough to enable statistics reporting, it is\n recommended to set all other settings in order to avoid relying on default\n unobvious parameters.\n\n Example :\n # public access (limited to this backend only)\n backend public_www\n server srv1 192.168.0.1:80\n stats enable\n stats hide-version\n stats scope .\n stats uri /admin?stats\n stats realm Haproxy\ Statistics\n stats auth admin1:AdMiN123\n stats auth admin2:AdMiN321\n\n # internal monitoring access (unlimited)\n backend private_monitoring\n stats enable\n stats uri /admin?stats\n stats refresh 5s\n\n See also : "stats auth", "stats realm", "stats uri"\n\n\nstats hide-version\n Enable statistics and hide HAProxy version reporting\n May be used in sections : defaults | frontend | listen | backend\n yes | no | yes | yes\n Arguments : none\n\n By default, the stats page reports some useful status information along with\n the statistics. Among them is HAProxy's version. However, it is generally\n considered dangerous to report precise version to anyone, as it can help them\n target known weaknesses with specific attacks. The "stats hide-version"\n statement removes the version from the statistics report. This is recommended\n for public sites or any site with a weak login/password.\n\n Though this statement alone is enough to enable statistics reporting, it is\n recommended to set all other settings in order to avoid relying on default\n unobvious parameters.\n\n Example :\n # public access (limited to this backend only)\n backend public_www\n server srv1 192.168.0.1:80\n stats enable\n stats hide-version\n stats scope .\n stats uri /admin?stats\n stats realm Haproxy\ Statistics\n stats auth admin1:AdMiN123\n stats auth admin2:AdMiN321\n\n # internal monitoring access (unlimited)\n backend private_monitoring\n stats enable\n stats uri /admin?stats\n stats refresh 5s\n\n See also : "stats auth", "stats enable", "stats realm", "stats uri"\n\n\nstats http-request { allow | deny | auth [realm <realm>] }\n [ { if | unless } <condition> ]\n Access control for statistics\n\n May be used in sections: defaults | frontend | listen | backend\n no | no | yes | yes\n\n As "http-request", these set of options allow to fine control access to\n statistics. Each option may be followed by if/unless and acl.\n First option with matched condition (or option without condition) is final.\n For "deny" a 403 error will be returned, for "allow" normal processing is\n performed, for "auth" a 401/407 error code is returned so the client\n should be asked to enter a username and password.\n\n There is no fixed limit to the number of http-request statements per\n instance.\n\n See also : "http-request", section 3.4 about userlists and section 7\n about ACL usage.\n\n\nstats realm <realm>\n Enable statistics and set authentication realm\n May be used in sections : defaults | frontend | listen | backend\n yes | no | yes | yes\n Arguments :\n <realm> is the name of the HTTP Basic Authentication realm reported to\n the browser. The browser uses it to display it in the pop-up\n inviting the user to enter a valid username and password.\n\n The realm is read as a single word, so any spaces in it should be escaped\n using a backslash ('\').\n\n This statement is useful only in conjunction with "stats auth" since it is\n only related to authentication.\n\n Though this statement alone is enough to enable statistics reporting, it is\n recommended to set all other settings in order to avoid relying on default\n unobvious parameters.\n\n Example :\n # public access (limited to this backend only)\n backend public_www\n server srv1 192.168.0.1:80\n stats enable\n stats hide-version\n stats scope .\n stats uri /admin?stats\n stats realm Haproxy\ Statistics\n stats auth admin1:AdMiN123\n stats auth admin2:AdMiN321\n\n # internal monitoring access (unlimited)\n backend private_monitoring\n stats enable\n stats uri /admin?stats\n stats refresh 5s\n\n See also : "stats auth", "stats enable", "stats uri"\n\n\nstats refresh <delay>\n Enable statistics with automatic refresh\n May be used in sections : defaults | frontend | listen | backend\n yes | no | yes | yes\n Arguments :\n <delay> is the suggested refresh delay, specified in seconds, which will\n be returned to the browser consulting the report page. While the\n browser is free to apply any delay, it will generally respect it\n and refresh the page this every seconds. The refresh interval may\n be specified in any other non-default time unit, by suffixing the\n unit after the value, as explained at the top of this document.\n\n This statement is useful on monitoring displays with a permanent page\n reporting the load balancer's activity. When set, the HTML report page will\n include a link "refresh"/"stop refresh" so that the user can select whether\n he wants automatic refresh of the page or not.\n\n Though this statement alone is enough to enable statistics reporting, it is\n recommended to set all other settings in order to avoid relying on default\n unobvious parameters.\n\n Example :\n # public access (limited to this backend only)\n backend public_www\n server srv1 192.168.0.1:80\n stats enable\n stats hide-version\n stats scope .\n stats uri /admin?stats\n stats realm Haproxy\ Statistics\n stats auth admin1:AdMiN123\n stats auth admin2:AdMiN321\n\n # internal monitoring access (unlimited)\n backend private_monitoring\n stats enable\n stats uri /admin?stats\n stats refresh 5s\n\n See also : "stats auth", "stats enable", "stats realm", "stats uri"\n\n\nstats scope { <name> | "." }\n Enable statistics and limit access scope\n May be used in sections : defaults | frontend | listen | backend\n yes | no | yes | yes\n Arguments :\n <name> is the name of a listen, frontend or backend section to be\n reported. The special name "." (a single dot) designates the\n section in which the statement appears.\n\n When this statement is specified, only the sections enumerated with this\n statement will appear in the report. All other ones will be hidden. This\n statement may appear as many times as needed if multiple sections need to be\n reported. Please note that the name checking is performed as simple string\n comparisons, and that it is never checked that a give section name really\n exists.\n\n Though this statement alone is enough to enable statistics reporting, it is\n recommended to set all other settings in order to avoid relying on default\n unobvious parameters.\n\n Example :\n # public access (limited to this backend only)\n backend public_www\n server srv1 192.168.0.1:80\n stats enable\n stats hide-version\n stats scope .\n stats uri /admin?stats\n stats realm Haproxy\ Statistics\n stats auth admin1:AdMiN123\n stats auth admin2:AdMiN321\n\n # internal monitoring access (unlimited)\n backend private_monitoring\n stats enable\n stats uri /admin?stats\n stats refresh 5s\n\n See also : "stats auth", "stats enable", "stats realm", "stats uri"\n\n\nstats show-desc [ <description> ]\n Enable reporting of a description on the statistics page.\n May be used in sections : defaults | frontend | listen | backend\n yes | no | yes | yes\n\n <name> is an optional description to be reported. If unspecified, the\n description from global section is automatically used instead.\n\n This statement is useful for users that offer shared services to their\n customers, where node or description should be different for each customer.\n\n Though this statement alone is enough to enable statistics reporting, it is\n recommended to set all other settings in order to avoid relying on default\n unobvious parameters. By default description is not shown.\n\n Example :\n # internal monitoring access (unlimited)\n backend private_monitoring\n stats enable\n stats show-desc Master node for Europe, Asia, Africa\n stats uri /admin?stats\n stats refresh 5s\n\n See also: "show-node", "stats enable", "stats uri" and "description" in\n global section.\n\n\nstats show-legends\n Enable reporting additional informations on the statistics page :\n - cap: capabilities (proxy)\n - mode: one of tcp, http or health (proxy)\n - id: SNMP ID (proxy, socket, server)\n - IP (socket, server)\n - cookie (backend, server)\n\n Though this statement alone is enough to enable statistics reporting, it is\n recommended to set all other settings in order to avoid relying on default\n unobvious parameters. Default behaviour is not to show this information.\n\n See also: "stats enable", "stats uri".\n\n\nstats show-node [ <name> ]\n Enable reporting of a host name on the statistics page.\n May be used in sections : defaults | frontend | listen | backend\n yes | no | yes | yes\n Arguments:\n <name> is an optional name to be reported. If unspecified, the\n node name from global section is automatically used instead.\n\n This statement is useful for users that offer shared services to their\n customers, where node or description might be different on a stats page\n provided for each customer. Default behaviour is not to show host name.\n\n Though this statement alone is enough to enable statistics reporting, it is\n recommended to set all other settings in order to avoid relying on default\n unobvious parameters.\n\n Example:\n # internal monitoring access (unlimited)\n backend private_monitoring\n stats enable\n stats show-node Europe-1\n stats uri /admin?stats\n stats refresh 5s\n\n See also: "show-desc", "stats enable", "stats uri", and "node" in global\n section.\n\n\nstats uri <prefix>\n Enable statistics and define the URI prefix to access them\n May be used in sections : defaults | frontend | listen | backend\n yes | no | yes | yes\n Arguments :\n <prefix> is the prefix of any URI which will be redirected to stats. This\n prefix may contain a question mark ('?') to indicate part of a\n query string.\n\n The statistics URI is intercepted on the relayed traffic, so it appears as a\n page within the normal application. It is strongly advised to ensure that the\n selected URI will never appear in the application, otherwise it will never be\n possible to reach it in the application.\n\n The default URI compiled in haproxy is "/haproxy?stats", but this may be\n changed at build time, so it's better to always explicitly specify it here.\n It is generally a good idea to include a question mark in the URI so that\n intermediate proxies refrain from caching the results. Also, since any string\n beginning with the prefix will be accepted as a stats request, the question\n mark helps ensuring that no valid URI will begin with the same words.\n\n It is sometimes very convenient to use "/" as the URI prefix, and put that\n statement in a "listen" instance of its own. That makes it easy to dedicate\n an address or a port to statistics only.\n\n Though this statement alone is enough to enable statistics reporting, it is\n recommended to set all other settings in order to avoid relying on default\n unobvious parameters.\n\n Example :\n # public access (limited to this backend only)\n backend public_www\n server srv1 192.168.0.1:80\n stats enable\n stats hide-version\n stats scope .\n stats uri /admin?stats\n stats realm Haproxy\ Statistics\n stats auth admin1:AdMiN123\n stats auth admin2:AdMiN321\n\n # internal monitoring access (unlimited)\n backend private_monitoring\n stats enable\n stats uri /admin?stats\n stats refresh 5s\n\n See also : "stats auth", "stats enable", "stats realm"\n\n\nstick match <pattern> [table <table>] [{if | unless} <cond>]\n Define a request pattern matching condition to stick a user to a server\n May be used in sections : defaults | frontend | listen | backend\n no | no | yes | yes\n\n Arguments :\n <pattern> is a pattern extraction rule as described in section 7.8. It\n describes what elements of the incoming request or connection\n will be analysed in the hope to find a matching entry in a\n stickiness table. This rule is mandatory.\n\n <table> is an optional stickiness table name. If unspecified, the same\n backend's table is used. A stickiness table is declared using\n the "stick-table" statement.\n\n <cond> is an optional matching condition. It makes it possible to match\n on a certain criterion only when other conditions are met (or\n not met). For instance, it could be used to match on a source IP\n address except when a request passes through a known proxy, in\n which case we'd match on a header containing that IP address.\n\n Some protocols or applications require complex stickiness rules and cannot\n always simply rely on cookies nor hashing. The "stick match" statement\n describes a rule to extract the stickiness criterion from an incoming request\n or connection. See section 7 for a complete list of possible patterns and\n transformation rules.\n\n The table has to be declared using the "stick-table" statement. It must be of\n a type compatible with the pattern. By default it is the one which is present\n in the same backend. It is possible to share a table with other backends by\n referencing it using the "table" keyword. If another table is referenced,\n the server's ID inside the backends are used. By default, all server IDs\n start at 1 in each backend, so the server ordering is enough. But in case of\n doubt, it is highly recommended to force server IDs using their "id" setting.\n\n It is possible to restrict the conditions where a "stick match" statement\n will apply, using "if" or "unless" followed by a condition. See section 7 for\n ACL based conditions.\n\n There is no limit on the number of "stick match" statements. The first that\n applies and matches will cause the request to be directed to the same server\n as was used for the request which created the entry. That way, multiple\n matches can be used as fallbacks.\n\n The stick rules are checked after the persistence cookies, so they will not\n affect stickiness if a cookie has already been used to select a server. That\n way, it becomes very easy to insert cookies and match on IP addresses in\n order to maintain stickiness between HTTP and HTTPS.\n\n Note : Consider not using this feature in multi-process mode (nbproc > 1)\n unless you know what you do : memory is not shared between the\n processes, which can result in random behaviours.\n\n Example :\n # forward SMTP users to the same server they just used for POP in the\n # last 30 minutes\n backend pop\n mode tcp\n balance roundrobin\n stick store-request src\n stick-table type ip size 200k expire 30m\n server s1 192.168.1.1:110\n server s2 192.168.1.1:110\n\n backend smtp\n mode tcp\n balance roundrobin\n stick match src table pop\n server s1 192.168.1.1:25\n server s2 192.168.1.1:25\n\n See also : "stick-table", "stick on", "nbproc", "bind-process" and section 7\n about ACLs and pattern extraction.\n\n\nstick on <pattern> [table <table>] [{if | unless} <condition>]\n Define a request pattern to associate a user to a server\n May be used in sections : defaults | frontend | listen | backend\n no | no | yes | yes\n\n Note : This form is exactly equivalent to "stick match" followed by\n "stick store-request", all with the same arguments. Please refer\n to both keywords for details. It is only provided as a convenience\n for writing more maintainable configurations.\n\n Note : Consider not using this feature in multi-process mode (nbproc > 1)\n unless you know what you do : memory is not shared between the\n processes, which can result in random behaviours.\n\n Examples :\n # The following form ...\n stick on src table pop if !localhost\n\n # ...is strictly equivalent to this one :\n stick match src table pop if !localhost\n stick store-request src table pop if !localhost\n\n\n # Use cookie persistence for HTTP, and stick on source address for HTTPS as\n # well as HTTP without cookie. Share the same table between both accesses.\n backend http\n mode http\n balance roundrobin\n stick on src table https\n cookie SRV insert indirect nocache\n server s1 192.168.1.1:80 cookie s1\n server s2 192.168.1.1:80 cookie s2\n\n backend https\n mode tcp\n balance roundrobin\n stick-table type ip size 200k expire 30m\n stick on src\n server s1 192.168.1.1:443\n server s2 192.168.1.1:443\n\n See also : "stick match", "stick store-request", "nbproc" and "bind-process".\n\n\nstick store-request <pattern> [table <table>] [{if | unless} <condition>]\n Define a request pattern used to create an entry in a stickiness table\n May be used in sections : defaults | frontend | listen | backend\n no | no | yes | yes\n\n Arguments :\n <pattern> is a pattern extraction rule as described in section 7.8. It\n describes what elements of the incoming request or connection\n will be analysed, extracted and stored in the table once a\n server is selected.\n\n <table> is an optional stickiness table name. If unspecified, the same\n backend's table is used. A stickiness table is declared using\n the "stick-table" statement.\n\n <cond> is an optional storage condition. It makes it possible to store\n certain criteria only when some conditions are met (or not met).\n For instance, it could be used to store the source IP address\n except when the request passes through a known proxy, in which\n case we'd store a converted form of a header containing that IP\n address.\n\n Some protocols or applications require complex stickiness rules and cannot\n always simply rely on cookies nor hashing. The "stick store-request" statement\n describes a rule to decide what to extract from the request and when to do\n it, in order to store it into a stickiness table for further requests to\n match it using the "stick match" statement. Obviously the extracted part must\n make sense and have a chance to be matched in a further request. Storing a\n client's IP address for instance often makes sense. Storing an ID found in a\n URL parameter also makes sense. Storing a source port will almost never make\n any sense because it will be randomly matched. See section 7 for a complete\n list of possible patterns and transformation rules.\n\n The table has to be declared using the "stick-table" statement. It must be of\n a type compatible with the pattern. By default it is the one which is present\n in the same backend. It is possible to share a table with other backends by\n referencing it using the "table" keyword. If another table is referenced,\n the server's ID inside the backends are used. By default, all server IDs\n start at 1 in each backend, so the server ordering is enough. But in case of\n doubt, it is highly recommended to force server IDs using their "id" setting.\n\n It is possible to restrict the conditions where a "stick store-request"\n statement will apply, using "if" or "unless" followed by a condition. This\n condition will be evaluated while parsing the request, so any criteria can be\n used. See section 7 for ACL based conditions.\n\n There is no limit on the number of "stick store-request" statements, but\n there is a limit of 8 simultaneous stores per request or response. This\n makes it possible to store up to 8 criteria, all extracted from either the\n request or the response, regardless of the number of rules. Only the 8 first\n ones which match will be kept. Using this, it is possible to feed multiple\n tables at once in the hope to increase the chance to recognize a user on\n another protocol or access method.\n\n The "store-request" rules are evaluated once the server connection has been\n established, so that the table will contain the real server that processed\n the request.\n\n Note : Consider not using this feature in multi-process mode (nbproc > 1)\n unless you know what you do : memory is not shared between the\n processes, which can result in random behaviours.\n\n Example :\n # forward SMTP users to the same server they just used for POP in the\n # last 30 minutes\n backend pop\n mode tcp\n balance roundrobin\n stick store-request src\n stick-table type ip size 200k expire 30m\n server s1 192.168.1.1:110\n server s2 192.168.1.1:110\n\n backend smtp\n mode tcp\n balance roundrobin\n stick match src table pop\n server s1 192.168.1.1:25\n server s2 192.168.1.1:25\n\n See also : "stick-table", "stick on", "nbproc", "bind-process" and section 7\n about ACLs and pattern extraction.\n\n\nstick-table type {ip | integer | string [len <length>] } size <size>\n [expire <expire>] [nopurge]\n Configure the stickiness table for the current backend\n May be used in sections : defaults | frontend | listen | backend\n no | no | yes | yes\n\n Arguments :\n ip a table declared with "type ip" will only store IPv4 addresses.\n This form is very compact (about 50 bytes per entry) and allows\n very fast entry lookup and stores with almost no overhead. This\n is mainly used to store client source IP addresses.\n\n integer a table declared with "type integer" will store 32bit integers\n which can represent a client identifier found in a request for\n instance.\n\n string a table declared with "type string" will store substrings of up\n to <len> characters. If the string provided by the pattern\n extractor is larger than <len>, it will be truncated before\n being stored. During matching, at most <len> characters will be\n compared between the string in the table and the extracted\n pattern. When not specified, the string is automatically limited\n to 31 characters.\n\n <length> is the maximum number of characters that will be stored in a\n "string" type table. See type "string" above. Be careful when\n changing this parameter as memory usage will proportionally\n increase.\n\n <size> is the maximum number of entries that can fit in the table. This\n value directly impacts memory usage. Count approximately\n 50 bytes per entry, plus the size of a string if any. The size\n supports suffixes "k", "m", "g" for 2^10, 2^20 and 2^30 factors.\n\n [nopurge] indicates that we refuse to purge older entries when the table\n is full. When not specified and the table is full when haproxy\n wants to store an entry in it, it will flush a few of the oldest\n entries in order to release some space for the new ones. This is\n most often the desired behaviour. In some specific cases, it\n be desirable to refuse new entries instead of purging the older\n ones. That may be the case when the amount of data to store is\n far above the hardware limits and we prefer not to offer access\n to new clients than to reject the ones already connected. When\n using this parameter, be sure to properly set the "expire"\n parameter (see below).\n\n <expire> defines the maximum duration of an entry in the table since it\n was last created, refreshed or matched. The expiration delay is\n defined using the standard time format, similarly as the various\n timeouts. The maximum duration is slightly above 24 days. See\n section 2.2 for more information. If this delay is not specified,\n the session won't automatically expire, but older entries will\n be removed once full. Be sure not to use the "nopurge" parameter\n if not expiration delay is specified.\n\n The is only one stick-table per backend. At the moment of writing this doc,\n it does not seem useful to have multiple tables per backend. If this happens\n to be required, simply create a dummy backend with a stick-table in it and\n reference it.\n\n It is important to understand that stickiness based on learning information\n has some limitations, including the fact that all learned associations are\n lost upon restart. In general it can be good as a complement but not always\n as an exclusive stickiness.\n\n See also : "stick match", "stick on", "stick store-request", and section 2.2\n about time format.\n\n\ntcp-request content accept [{if | unless} <condition>]\n Accept a connection if/unless a content inspection condition is matched\n May be used in sections : defaults | frontend | listen | backend\n no | yes | yes | no\n\n During TCP content inspection, the connection is immediately validated if the\n condition is true (when used with "if") or false (when used with "unless").\n Most of the time during content inspection, a condition will be in an\n uncertain state which is neither true nor false. The evaluation immediately\n stops when such a condition is encountered. It is important to understand\n that "accept" and "reject" rules are evaluated in their exact declaration\n order, so that it is possible to build complex rules from them. There is no\n specific limit to the number of rules which may be inserted.\n\n Note that the "if/unless" condition is optional. If no condition is set on\n the action, it is simply performed unconditionally.\n\n If no "tcp-request content" rules are matched, the default action already is\n "accept". Thus, this statement alone does not bring anything without another\n "reject" statement.\n\n See section 7 about ACL usage.\n\n See also : "tcp-request content reject", "tcp-request inspect-delay"\n\n\ntcp-request content reject [{if | unless} <condition>]\n Reject a connection if/unless a content inspection condition is matched\n May be used in sections : defaults | frontend | listen | backend\n no | yes | yes | no\n\n During TCP content inspection, the connection is immediately rejected if the\n condition is true (when used with "if") or false (when used with "unless").\n Most of the time during content inspection, a condition will be in an\n uncertain state which is neither true nor false. The evaluation immediately\n stops when such a condition is encountered. It is important to understand\n that "accept" and "reject" rules are evaluated in their exact declaration\n order, so that it is possible to build complex rules from them. There is no\n specific limit to the number of rules which may be inserted.\n\n Note that the "if/unless" condition is optional. If no condition is set on\n the action, it is simply performed unconditionally.\n\n If no "tcp-request content" rules are matched, the default action is set to\n "accept".\n\n Example:\n # reject SMTP connection if client speaks first\n tcp-request inspect-delay 30s\n acl content_present req_len gt 0\n tcp-request reject if content_present\n\n # Forward HTTPS connection only if client speaks\n tcp-request inspect-delay 30s\n acl content_present req_len gt 0\n tcp-request accept if content_present\n tcp-request reject\n\n See section 7 about ACL usage.\n\n See also : "tcp-request content accept", "tcp-request inspect-delay"\n\n\ntcp-request inspect-delay <timeout>\n Set the maximum allowed time to wait for data during content inspection\n May be used in sections : defaults | frontend | listen | backend\n no | yes | yes | no\n Arguments :\n <timeout> is the timeout value specified in milliseconds by default, but\n can be in any other unit if the number is suffixed by the unit,\n as explained at the top of this document.\n\n People using haproxy primarily as a TCP relay are often worried about the\n risk of passing any type of protocol to a server without any analysis. In\n order to be able to analyze the request contents, we must first withhold\n the data then analyze them. This statement simply enables withholding of\n data for at most the specified amount of time.\n\n Note that when performing content inspection, haproxy will evaluate the whole\n rules for every new chunk which gets in, taking into account the fact that\n those data are partial. If no rule matches before the aforementioned delay,\n a last check is performed upon expiration, this time considering that the\n contents are definitive. If no delay is set, haproxy will not wait at all\n and will immediately apply a verdict based on the available information.\n Obviously this is unlikely to be very useful and might even be racy, so such\n setups are not recommended.\n\n As soon as a rule matches, the request is released and continues as usual. If\n the timeout is reached and no rule matches, the default policy will be to let\n it pass through unaffected.\n\n For most protocols, it is enough to set it to a few seconds, as most clients\n send the full request immediately upon connection. Add 3 or more seconds to\n cover TCP retransmits but that's all. For some protocols, it may make sense\n to use large values, for instance to ensure that the client never talks\n before the server (eg: SMTP), or to wait for a client to talk before passing\n data to the server (eg: SSL). Note that the client timeout must cover at\n least the inspection delay, otherwise it will expire first. If the client\n closes the connection or if the buffer is full, the delay immediately expires\n since the contents will not be able to change anymore.\n\n See also : "tcp-request content accept", "tcp-request content reject",\n "timeout client".\n\n\ntimeout check <timeout>\n Set additional check timeout, but only after a connection has been already\n established.\n\n May be used in sections: defaults | frontend | listen | backend\n yes | no | yes | yes\n Arguments:\n <timeout> is the timeout value specified in milliseconds by default, but\n can be in any other unit if the number is suffixed by the unit,\n as explained at the top of this document.\n\n If set, haproxy uses min("timeout connect", "inter") as a connect timeout\n for check and "timeout check" as an additional read timeout. The "min" is\n used so that people running with *very* long "timeout connect" (eg. those\n who needed this due to the queue or tarpit) do not slow down their checks.\n (Please also note that there is no valid reason to have such long connect\n timeouts, because "timeout queue" and "timeout tarpit" can always be used to\n avoid that).\n\n If "timeout check" is not set haproxy uses "inter" for complete check\n timeout (connect + read) exactly like all <1.3.15 version.\n\n In most cases check request is much simpler and faster to handle than normal\n requests and people may want to kick out laggy servers so this timeout should\n be smaller than "timeout server".\n\n This parameter is specific to backends, but can be specified once for all in\n "defaults" sections. This is in fact one of the easiest solutions not to\n forget about it.\n\n See also: "timeout connect", "timeout queue", "timeout server",\n "timeout tarpit".\n\n\ntimeout client <timeout>\ntimeout clitimeout <timeout> (deprecated)\n Set the maximum inactivity time on the client side.\n May be used in sections : defaults | frontend | listen | backend\n yes | yes | yes | no\n Arguments :\n <timeout> is the timeout value specified in milliseconds by default, but\n can be in any other unit if the number is suffixed by the unit,\n as explained at the top of this document.\n\n The inactivity timeout applies when the client is expected to acknowledge or\n send data. In HTTP mode, this timeout is particularly important to consider\n during the first phase, when the client sends the request, and during the\n response while it is reading data sent by the server. The value is specified\n in milliseconds by default, but can be in any other unit if the number is\n suffixed by the unit, as specified at the top of this document. In TCP mode\n (and to a lesser extent, in HTTP mode), it is highly recommended that the\n client timeout remains equal to the server timeout in order to avoid complex\n situations to debug. It is a good practice to cover one or several TCP packet\n losses by specifying timeouts that are slightly above multiples of 3 seconds\n (eg: 4 or 5 seconds).\n\n This parameter is specific to frontends, but can be specified once for all in\n "defaults" sections. This is in fact one of the easiest solutions not to\n forget about it. An unspecified timeout results in an infinite timeout, which\n is not recommended. Such a usage is accepted and works but reports a warning\n during startup because it may results in accumulation of expired sessions in\n the system if the system's timeouts are not configured either.\n\n This parameter replaces the old, deprecated "clitimeout". It is recommended\n to use it to write new configurations. The form "timeout clitimeout" is\n provided only by backwards compatibility but its use is strongly discouraged.\n\n See also : "clitimeout", "timeout server".\n\n\ntimeout connect <timeout>\ntimeout contimeout <timeout> (deprecated)\n Set the maximum time to wait for a connection attempt to a server to succeed.\n May be used in sections : defaults | frontend | listen | backend\n yes | no | yes | yes\n Arguments :\n <timeout> is the timeout value specified in milliseconds by default, but\n can be in any other unit if the number is suffixed by the unit,\n as explained at the top of this document.\n\n If the server is located on the same LAN as haproxy, the connection should be\n immediate (less than a few milliseconds). Anyway, it is a good practice to\n cover one or several TCP packet losses by specifying timeouts that are\n slightly above multiples of 3 seconds (eg: 4 or 5 seconds). By default, the\n connect timeout also presets both queue and tarpit timeouts to the same value\n if these have not been specified.\n\n This parameter is specific to backends, but can be specified once for all in\n "defaults" sections. This is in fact one of the easiest solutions not to\n forget about it. An unspecified timeout results in an infinite timeout, which\n is not recommended. Such a usage is accepted and works but reports a warning\n during startup because it may results in accumulation of failed sessions in\n the system if the system's timeouts are not configured either.\n\n This parameter replaces the old, deprecated "contimeout". It is recommended\n to use it to write new configurations. The form "timeout contimeout" is\n provided only by backwards compatibility but its use is strongly discouraged.\n\n See also: "timeout check", "timeout queue", "timeout server", "contimeout",\n "timeout tarpit".\n\n\ntimeout http-keep-alive <timeout>\n Set the maximum allowed time to wait for a new HTTP request to appear\n May be used in sections : defaults | frontend | listen | backend\n yes | yes | yes | yes\n Arguments :\n <timeout> is the timeout value specified in milliseconds by default, but\n can be in any other unit if the number is suffixed by the unit,\n as explained at the top of this document.\n\n By default, the time to wait for a new request in case of keep-alive is set\n by "timeout http-request". However this is not always convenient because some\n people want very short keep-alive timeouts in order to release connections\n faster, and others prefer to have larger ones but still have short timeouts\n once the request has started to present itself.\n\n The "http-keep-alive" timeout covers these needs. It will define how long to\n wait for a new HTTP request to start coming after a response was sent. Once\n the first byte of request has been seen, the "http-request" timeout is used\n to wait for the complete request to come. Note that empty lines prior to a\n new request do not refresh the timeout and are not counted as a new request.\n\n There is also another difference between the two timeouts : when a connection\n expires during timeout http-keep-alive, no error is returned, the connection\n just closes. If the connection expires in "http-request" while waiting for a\n connection to complete, a HTTP 408 error is returned.\n\n In general it is optimal to set this value to a few tens to hundreds of\n milliseconds, to allow users to fetch all objects of a page at once but\n without waiting for further clicks. Also, if set to a very small value (eg:\n 1 millisecond) it will probably only accept pipelined requests but not the\n non-pipelined ones. It may be a nice trade-off for very large sites running\n with tens to hundreds of thousands of clients.\n\n If this parameter is not set, the "http-request" timeout applies, and if both\n are not set, "timeout client" still applies at the lower level. It should be\n set in the frontend to take effect, unless the frontend is in TCP mode, in\n which case the HTTP backend's timeout will be used.\n\n See also : "timeout http-request", "timeout client".\n\n\ntimeout http-request <timeout>\n Set the maximum allowed time to wait for a complete HTTP request\n May be used in sections : defaults | frontend | listen | backend\n yes | yes | yes | yes\n Arguments :\n <timeout> is the timeout value specified in milliseconds by default, but\n can be in any other unit if the number is suffixed by the unit,\n as explained at the top of this document.\n\n In order to offer DoS protection, it may be required to lower the maximum\n accepted time to receive a complete HTTP request without affecting the client\n timeout. This helps protecting against established connections on which\n nothing is sent. The client timeout cannot offer a good protection against\n this abuse because it is an inactivity timeout, which means that if the\n attacker sends one character every now and then, the timeout will not\n trigger. With the HTTP request timeout, no matter what speed the client\n types, the request will be aborted if it does not complete in time.\n\n Note that this timeout only applies to the header part of the request, and\n not to any data. As soon as the empty line is received, this timeout is not\n used anymore. It is used again on keep-alive connections to wait for a second\n request if "timeout http-keep-alive" is not set.\n\n Generally it is enough to set it to a few seconds, as most clients send the\n full request immediately upon connection. Add 3 or more seconds to cover TCP\n retransmits but that's all. Setting it to very low values (eg: 50 ms) will\n generally work on local networks as long as there are no packet losses. This\n will prevent people from sending bare HTTP requests using telnet.\n\n If this parameter is not set, the client timeout still applies between each\n chunk of the incoming request. It should be set in the frontend to take\n effect, unless the frontend is in TCP mode, in which case the HTTP backend's\n timeout will be used.\n\n See also : "timeout http-keep-alive", "timeout client".\n\n\ntimeout queue <timeout>\n Set the maximum time to wait in the queue for a connection slot to be free\n May be used in sections : defaults | frontend | listen | backend\n yes | no | yes | yes\n Arguments :\n <timeout> is the timeout value specified in milliseconds by default, but\n can be in any other unit if the number is suffixed by the unit,\n as explained at the top of this document.\n\n When a server's maxconn is reached, connections are left pending in a queue\n which may be server-specific or global to the backend. In order not to wait\n indefinitely, a timeout is applied to requests pending in the queue. If the\n timeout is reached, it is considered that the request will almost never be\n served, so it is dropped and a 503 error is returned to the client.\n\n The "timeout queue" statement allows to fix the maximum time for a request to\n be left pending in a queue. If unspecified, the same value as the backend's\n connection timeout ("timeout connect") is used, for backwards compatibility\n with older versions with no "timeout queue" parameter.\n\n See also : "timeout connect", "contimeout".\n\n\ntimeout server <timeout>\ntimeout srvtimeout <timeout> (deprecated)\n Set the maximum inactivity time on the server side.\n May be used in sections : defaults | frontend | listen | backend\n yes | no | yes | yes\n Arguments :\n <timeout> is the timeout value specified in milliseconds by default, but\n can be in any other unit if the number is suffixed by the unit,\n as explained at the top of this document.\n\n The inactivity timeout applies when the server is expected to acknowledge or\n send data. In HTTP mode, this timeout is particularly important to consider\n during the first phase of the server's response, when it has to send the\n headers, as it directly represents the server's processing time for the\n request. To find out what value to put there, it's often good to start with\n what would be considered as unacceptable response times, then check the logs\n to observe the response time distribution, and adjust the value accordingly.\n\n The value is specified in milliseconds by default, but can be in any other\n unit if the number is suffixed by the unit, as specified at the top of this\n document. In TCP mode (and to a lesser extent, in HTTP mode), it is highly\n recommended that the client timeout remains equal to the server timeout in\n order to avoid complex situations to debug. Whatever the expected server\n response times, it is a good practice to cover at least one or several TCP\n packet losses by specifying timeouts that are slightly above multiples of 3\n seconds (eg: 4 or 5 seconds minimum).\n\n This parameter is specific to backends, but can be specified once for all in\n "defaults" sections. This is in fact one of the easiest solutions not to\n forget about it. An unspecified timeout results in an infinite timeout, which\n is not recommended. Such a usage is accepted and works but reports a warning\n during startup because it may results in accumulation of expired sessions in\n the system if the system's timeouts are not configured either.\n\n This parameter replaces the old, deprecated "srvtimeout". It is recommended\n to use it to write new configurations. The form "timeout srvtimeout" is\n provided only by backwards compatibility but its use is strongly discouraged.\n\n See also : "srvtimeout", "timeout client".\n\n\ntimeout tarpit <timeout>\n Set the duration for which tarpitted connections will be maintained\n May be used in sections : defaults | frontend | listen | backend\n yes | yes | yes | yes\n Arguments :\n <timeout> is the tarpit duration specified in milliseconds by default, but\n can be in any other unit if the number is suffixed by the unit,\n as explained at the top of this document.\n\n When a connection is tarpitted using "reqtarpit", it is maintained open with\n no activity for a certain amount of time, then closed. "timeout tarpit"\n defines how long it will be maintained open.\n\n The value is specified in milliseconds by default, but can be in any other\n unit if the number is suffixed by the unit, as specified at the top of this\n document. If unspecified, the same value as the backend's connection timeout\n ("timeout connect") is used, for backwards compatibility with older versions\n with no "timeout tarpit" parameter.\n\n See also : "timeout connect", "contimeout".\n\n\ntransparent (deprecated)\n Enable client-side transparent proxying\n May be used in sections : defaults | frontend | listen | backend\n yes | no | yes | yes\n Arguments : none\n\n This keyword was introduced in order to provide layer 7 persistence to layer\n 3 load balancers. The idea is to use the OS's ability to redirect an incoming\n connection for a remote address to a local process (here HAProxy), and let\n this process know what address was initially requested. When this option is\n used, sessions without cookies will be forwarded to the original destination\n IP address of the incoming request (which should match that of another\n equipment), while requests with cookies will still be forwarded to the\n appropriate server.\n\n The "transparent" keyword is deprecated, use "option transparent" instead.\n\n Note that contrary to a common belief, this option does NOT make HAProxy\n present the client's IP to the server when establishing the connection.\n\n See also: "option transparent"\n\n\nuse_backend <backend> if <condition>\nuse_backend <backend> unless <condition>\n Switch to a specific backend if/unless an ACL-based condition is matched.\n May be used in sections : defaults | frontend | listen | backend\n no | yes | yes | no\n Arguments :\n <backend> is the name of a valid backend or "listen" section.\n\n <condition> is a condition composed of ACLs, as described in section 7.\n\n When doing content-switching, connections arrive on a frontend and are then\n dispatched to various backends depending on a number of conditions. The\n relation between the conditions and the backends is described with the\n "use_backend" keyword. While it is normally used with HTTP processing, it can\n also be used in pure TCP, either without content using stateless ACLs (eg:\n source address validation) or combined with a "tcp-request" rule to wait for\n some payload.\n\n There may be as many "use_backend" rules as desired. All of these rules are\n evaluated in their declaration order, and the first one which matches will\n assign the backend.\n\n In the first form, the backend will be used if the condition is met. In the\n second form, the backend will be used if the condition is not met. If no\n condition is valid, the backend defined with "default_backend" will be used.\n If no default backend is defined, either the servers in the same section are\n used (in case of a "listen" section) or, in case of a frontend, no server is\n used and a 503 service unavailable response is returned.\n\n Note that it is possible to switch from a TCP frontend to an HTTP backend. In\n this case, either the frontend has already checked that the protocol is HTTP,\n and backend processing will immediately follow, or the backend will wait for\n a complete HTTP request to get in. This feature is useful when a frontend\n must decode several protocols on a unique port, one of them being HTTP.\n\n See also: "default_backend", "tcp-request", and section 7 about ACLs.\n\n\n5. Server and default-server options\n------------------------------------\n\nThe "server" and "default-server" keywords support a certain number of settings\nwhich are all passed as arguments on the server line. The order in which those\narguments appear does not count, and they are all optional. Some of those\nsettings are single words (booleans) while others expect one or several values\nafter them. In this case, the values must immediately follow the setting name.\nExcept default-server, all those settings must be specified after the server's\naddress if they are used:\n\n server <name> <address>[:port] [settings ...]\n default-server [settings ...]\n\nThe currently supported settings are the following ones.\n\naddr <ipv4>\n Using the "addr" parameter, it becomes possible to use a different IP address\n to send health-checks. On some servers, it may be desirable to dedicate an IP\n address to specific component able to perform complex tests which are more\n suitable to health-checks than the application. This parameter is ignored if\n the "check" parameter is not set. See also the "port" parameter.\n\n Supported in default-server: No\n\nbackup\n When "backup" is present on a server line, the server is only used in load\n balancing when all other non-backup servers are unavailable. Requests coming\n with a persistence cookie referencing the server will always be served\n though. By default, only the first operational backup server is used, unless\n the "allbackups" option is set in the backend. See also the "allbackups"\n option.\n\n Supported in default-server: No\n\ncheck\n This option enables health checks on the server. By default, a server is\n always considered available. If "check" is set, the server will receive\n periodic health checks to ensure that it is really able to serve requests.\n The default address and port to send the tests to are those of the server,\n and the default source is the same as the one defined in the backend. It is\n possible to change the address using the "addr" parameter, the port using the\n "port" parameter, the source address using the "source" address, and the\n interval and timers using the "inter", "rise" and "fall" parameters. The\n request method is define in the backend using the "httpchk", "smtpchk",\n "mysql-check" and "ssl-hello-chk" options. Please refer to those options and\n parameters for more information.\n\n Supported in default-server: No\n\ncookie <value>\n The "cookie" parameter sets the cookie value assigned to the server to\n <value>. This value will be checked in incoming requests, and the first\n operational server possessing the same value will be selected. In return, in\n cookie insertion or rewrite modes, this value will be assigned to the cookie\n sent to the client. There is nothing wrong in having several servers sharing\n the same cookie value, and it is in fact somewhat common between normal and\n backup servers. See also the "cookie" keyword in backend section.\n\n Supported in default-server: No\n\ndisabled\n The "disabled" keyword starts the server in the "disabled" state. That means\n that it is marked down in maintenance mode, and no connection other than the\n ones allowed by persist mode will reach it. It is very well suited to setup\n new servers, because normal traffic will never reach them, while it is still\n possible to test the service by making use of the force-persist mechanism.\n\n Supported in default-server: No\n\nerror-limit <count>\n If health observing is enabled, the "error-limit" parameter specifies the\n number of consecutive errors that triggers event selected by the "on-error"\n option. By default it is set to 10 consecutive errors.\n\n Supported in default-server: Yes\n\n See also the "check", "error-limit" and "on-error".\n\nfall <count>\n The "fall" parameter states that a server will be considered as dead after\n <count> consecutive unsuccessful health checks. This value defaults to 3 if\n unspecified. See also the "check", "inter" and "rise" parameters.\n\n Supported in default-server: Yes\n\nid <value>\n Set a persistent ID for the server. This ID must be positive and unique for\n the proxy. An unused ID will automatically be assigned if unset. The first\n assigned value will be 1. This ID is currently only returned in statistics.\n\n Supported in default-server: No\n\ninter <delay>\nfastinter <delay>\ndowninter <delay>\n The "inter" parameter sets the interval between two consecutive health checks\n to <delay> milliseconds. If left unspecified, the delay defaults to 2000 ms.\n It is also possible to use "fastinter" and "downinter" to optimize delays\n between checks depending on the server state :\n\n Server state | Interval used\n ---------------------------------+-----------------------------------------\n UP 100% (non-transitional) | "inter"\n ---------------------------------+-----------------------------------------\n Transitionally UP (going down), |\n Transitionally DOWN (going up), | "fastinter" if set, "inter" otherwise.\n or yet unchecked. |\n ---------------------------------+-----------------------------------------\n DOWN 100% (non-transitional) | "downinter" if set, "inter" otherwise.\n ---------------------------------+-----------------------------------------\n\n Just as with every other time-based parameter, they can be entered in any\n other explicit unit among { us, ms, s, m, h, d }. The "inter" parameter also\n serves as a timeout for health checks sent to servers if "timeout check" is\n not set. In order to reduce "resonance" effects when multiple servers are\n hosted on the same hardware, the health-checks of all servers are started\n with a small time offset between them. It is also possible to add some random\n noise in the health checks interval using the global "spread-checks"\n keyword. This makes sense for instance when a lot of backends use the same\n servers.\n\n Supported in default-server: Yes\n\nmaxconn <maxconn>\n The "maxconn" parameter specifies the maximal number of concurrent\n connections that will be sent to this server. If the number of incoming\n concurrent requests goes higher than this value, they will be queued, waiting\n for a connection to be released. This parameter is very important as it can\n save fragile servers from going down under extreme loads. If a "minconn"\n parameter is specified, the limit becomes dynamic. The default value is "0"\n which means unlimited. See also the "minconn" and "maxqueue" parameters, and\n the backend's "fullconn" keyword.\n\n Supported in default-server: Yes\n\nmaxqueue <maxqueue>\n The "maxqueue" parameter specifies the maximal number of connections which\n will wait in the queue for this server. If this limit is reached, next\n requests will be redispatched to other servers instead of indefinitely\n waiting to be served. This will break persistence but may allow people to\n quickly re-log in when the server they try to connect to is dying. The\n default value is "0" which means the queue is unlimited. See also the\n "maxconn" and "minconn" parameters.\n\n Supported in default-server: Yes\n\nminconn <minconn>\n When the "minconn" parameter is set, the maxconn limit becomes a dynamic\n limit following the backend's load. The server will always accept at least\n <minconn> connections, never more than <maxconn>, and the limit will be on\n the ramp between both values when the backend has less than <fullconn>\n concurrent connections. This makes it possible to limit the load on the\n server during normal loads, but push it further for important loads without\n overloading the server during exceptional loads. See also the "maxconn"\n and "maxqueue" parameters, as well as the "fullconn" backend keyword.\n\n Supported in default-server: Yes\n\nobserve <mode>\n This option enables health adjusting based on observing communication with\n the server. By default this functionality is disabled and enabling it also\n requires to enable health checks. There are two supported modes: "layer4" and\n "layer7". In layer4 mode, only successful/unsuccessful tcp connections are\n significant. In layer7, which is only allowed for http proxies, responses\n received from server are verified, like valid/wrong http code, unparsable\n headers, a timeout, etc. Valid status codes include 100 to 499, 501 and 505.\n\n Supported in default-server: No\n\n See also the "check", "on-error" and "error-limit".\n\non-error <mode>\n Select what should happen when enough consecutive errors are detected.\n Currently, four modes are available:\n - fastinter: force fastinter\n - fail-check: simulate a failed check, also forces fastinter (default)\n - sudden-death: simulate a pre-fatal failed health check, one more failed\n check will mark a server down, forces fastinter\n - mark-down: mark the server immediately down and force fastinter\n\n Supported in default-server: Yes\n\n See also the "check", "observe" and "error-limit".\n\nport <port>\n Using the "port" parameter, it becomes possible to use a different port to\n send health-checks. On some servers, it may be desirable to dedicate a port\n to a specific component able to perform complex tests which are more suitable\n to health-checks than the application. It is common to run a simple script in\n inetd for instance. This parameter is ignored if the "check" parameter is not\n set. See also the "addr" parameter.\n\n Supported in default-server: Yes\n\nredir <prefix>\n The "redir" parameter enables the redirection mode for all GET and HEAD\n requests addressing this server. This means that instead of having HAProxy\n forward the request to the server, it will send an "HTTP 302" response with\n the "Location" header composed of this prefix immediately followed by the\n requested URI beginning at the leading '/' of the path component. That means\n that no trailing slash should be used after <prefix>. All invalid requests\n will be rejected, and all non-GET or HEAD requests will be normally served by\n the server. Note that since the response is completely forged, no header\n mangling nor cookie insertion is possible in the response. However, cookies in\n requests are still analysed, making this solution completely usable to direct\n users to a remote location in case of local disaster. Main use consists in\n increasing bandwidth for static servers by having the clients directly\n connect to them. Note: never use a relative location here, it would cause a\n loop between the client and HAProxy!\n\n Example : server srv1 192.168.1.1:80 redir http://image1.mydomain.com check\n\n Supported in default-server: No\n\nrise <count>\n The "rise" parameter states that a server will be considered as operational\n after <count> consecutive successful health checks. This value defaults to 2\n if unspecified. See also the "check", "inter" and "fall" parameters.\n\n Supported in default-server: Yes\n\nslowstart <start_time_in_ms>\n The "slowstart" parameter for a server accepts a value in milliseconds which\n indicates after how long a server which has just come back up will run at\n full speed. Just as with every other time-based parameter, it can be entered\n in any other explicit unit among { us, ms, s, m, h, d }. The speed grows\n linearly from 0 to 100% during this time. The limitation applies to two\n parameters :\n\n - maxconn: the number of connections accepted by the server will grow from 1\n to 100% of the usual dynamic limit defined by (minconn,maxconn,fullconn).\n\n - weight: when the backend uses a dynamic weighted algorithm, the weight\n grows linearly from 1 to 100%. In this case, the weight is updated at every\n health-check. For this reason, it is important that the "inter" parameter\n is smaller than the "slowstart", in order to maximize the number of steps.\n\n The slowstart never applies when haproxy starts, otherwise it would cause\n trouble to running servers. It only applies when a server has been previously\n seen as failed.\n\n Supported in default-server: Yes\n\nsource <addr>[:<pl>[-<ph>]] [usesrc { <addr2>[:<port2>] | client | clientip } ]\nsource <addr>[:<port>] [usesrc { <addr2>[:<port2>] | hdr_ip(<hdr>[,<occ>]) } ]\nsource <addr>[:<pl>[-<ph>]] [interface <name>] ...\n The "source" parameter sets the source address which will be used when\n connecting to the server. It follows the exact same parameters and principle\n as the backend "source" keyword, except that it only applies to the server\n referencing it. Please consult the "source" keyword for details.\n\n Additionally, the "source" statement on a server line allows one to specify a\n source port range by indicating the lower and higher bounds delimited by a\n dash ('-'). Some operating systems might require a valid IP address when a\n source port range is specified. It is permitted to have the same IP/range for\n several servers. Doing so makes it possible to bypass the maximum of 64k\n total concurrent connections. The limit will then reach 64k connections per\n server.\n\n Supported in default-server: No\n\ntrack [<proxy>/]<server>\n This option enables ability to set the current state of the server by\n tracking another one. Only a server with checks enabled can be tracked\n so it is not possible for example to track a server that tracks another\n one. If <proxy> is omitted the current one is used. If disable-on-404 is\n used, it has to be enabled on both proxies.\n\n Supported in default-server: No\n\nweight <weight>\n The "weight" parameter is used to adjust the server's weight relative to\n other servers. All servers will receive a load proportional to their weight\n relative to the sum of all weights, so the higher the weight, the higher the\n load. The default weight is 1, and the maximal value is 256. A value of 0\n means the server will not participate in load-balancing but will still accept\n persistent connections. If this parameter is used to distribute the load\n according to server's capacity, it is recommended to start with values which\n can both grow and shrink, for instance between 10 and 100 to leave enough\n room above and below for later adjustments.\n\n Supported in default-server: Yes\n\n\n6. HTTP header manipulation\n---------------------------\n\nIn HTTP mode, it is possible to rewrite, add or delete some of the request and\nresponse headers based on regular expressions. It is also possible to block a\nrequest or a response if a particular header matches a regular expression,\nwhich is enough to stop most elementary protocol attacks, and to protect\nagainst information leak from the internal network. But there is a limitation\nto this : since HAProxy's HTTP engine does not support keep-alive, only headers\npassed during the first request of a TCP session will be seen. All subsequent\nheaders will be considered data only and not analyzed. Furthermore, HAProxy\nnever touches data contents, it stops analysis at the end of headers.\n\nThere is an exception though. If HAProxy encounters an "Informational Response"\n(status code 1xx), it is able to process all rsp* rules which can allow, deny,\nrewrite or delete a header, but it will refuse to add a header to any such\nmessages as this is not HTTP-compliant. The reason for still processing headers\nin such responses is to stop and/or fix any possible information leak which may\nhappen, for instance because another downstream equipment would unconditionally\nadd a header, or if a server name appears there. When such messages are seen,\nnormal processing still occurs on the next non-informational messages.\n\nThis section covers common usage of the following keywords, described in detail\nin section 4.2 :\n\n - reqadd <string>\n - reqallow <search>\n - reqiallow <search>\n - reqdel <search>\n - reqidel <search>\n - reqdeny <search>\n - reqideny <search>\n - reqpass <search>\n - reqipass <search>\n - reqrep <search> <replace>\n - reqirep <search> <replace>\n - reqtarpit <search>\n - reqitarpit <search>\n - rspadd <string>\n - rspdel <search>\n - rspidel <search>\n - rspdeny <search>\n - rspideny <search>\n - rsprep <search> <replace>\n - rspirep <search> <replace>\n\nWith all these keywords, the same conventions are used. The <search> parameter\nis a POSIX extended regular expression (regex) which supports grouping through\nparenthesis (without the backslash). Spaces and other delimiters must be\nprefixed with a backslash ('\') to avoid confusion with a field delimiter.\nOther characters may be prefixed with a backslash to change their meaning :\n\n \t for a tab\n \r for a carriage return (CR)\n \n for a new line (LF)\n \ to mark a space and differentiate it from a delimiter\n \# to mark a sharp and differentiate it from a comment\n \\ to use a backslash in a regex\n \\\\ to use a backslash in the text (*2 for regex, *2 for haproxy)\n \xXX to write the ASCII hex code XX as in the C language\n\nThe <replace> parameter contains the string to be used to replace the largest\nportion of text matching the regex. It can make use of the special characters\nabove, and can reference a substring which is delimited by parenthesis in the\nregex, by writing a backslash ('\') immediately followed by one digit from 0 to\n9 indicating the group position (0 designating the entire line). This practice\nis very common to users of the "sed" program.\n\nThe <string> parameter represents the string which will systematically be added\nafter the last header line. It can also use special character sequences above.\n\nNotes related to these keywords :\n---------------------------------\n - these keywords are not always convenient to allow/deny based on header\n contents. It is strongly recommended to use ACLs with the "block" keyword\n instead, resulting in far more flexible and manageable rules.\n\n - lines are always considered as a whole. It is not possible to reference\n a header name only or a value only. This is important because of the way\n headers are written (notably the number of spaces after the colon).\n\n - the first line is always considered as a header, which makes it possible to\n rewrite or filter HTTP requests URIs or response codes, but in turn makes\n it harder to distinguish between headers and request line. The regex prefix\n ^[^\ \t]*[\ \t] matches any HTTP method followed by a space, and the prefix\n ^[^ \t:]*: matches any header name followed by a colon.\n\n - for performances reasons, the number of characters added to a request or to\n a response is limited at build time to values between 1 and 4 kB. This\n should normally be far more than enough for most usages. If it is too short\n on occasional usages, it is possible to gain some space by removing some\n useless headers before adding new ones.\n\n - keywords beginning with "reqi" and "rspi" are the same as their counterpart\n without the 'i' letter except that they ignore case when matching patterns.\n\n - when a request passes through a frontend then a backend, all req* rules\n from the frontend will be evaluated, then all req* rules from the backend\n will be evaluated. The reverse path is applied to responses.\n\n - req* statements are applied after "block" statements, so that "block" is\n always the first one, but before "use_backend" in order to permit rewriting\n before switching.\n\n\n7. Using ACLs and pattern extraction\n------------------------------------\n\nThe use of Access Control Lists (ACL) provides a flexible solution to perform\ncontent switching and generally to take decisions based on content extracted\nfrom the request, the response or any environmental status. The principle is\nsimple :\n\n - define test criteria with sets of values\n - perform actions only if a set of tests is valid\n\nThe actions generally consist in blocking the request, or selecting a backend.\n\nIn order to define a test, the "acl" keyword is used. The syntax is :\n\n acl <aclname> <criterion> [flags] [operator] <value> ...\n\nThis creates a new ACL <aclname> or completes an existing one with new tests.\nThose tests apply to the portion of request/response specified in <criterion>\nand may be adjusted with optional flags [flags]. Some criteria also support\nan operator which may be specified before the set of values. The values are\nof the type supported by the criterion, and are separated by spaces.\n\nACL names must be formed from upper and lower case letters, digits, '-' (dash),\n'_' (underscore) , '.' (dot) and ':' (colon). ACL names are case-sensitive,\nwhich means that "my_acl" and "My_Acl" are two different ACLs.\n\nThere is no enforced limit to the number of ACLs. The unused ones do not affect\nperformance, they just consume a small amount of memory.\n\nThe following ACL flags are currently supported :\n\n -i : ignore case during matching of all subsequent patterns.\n -f : load patterns from a file.\n -- : force end of flags. Useful when a string looks like one of the flags.\n\nThe "-f" flag is special as it loads all of the lines it finds in the file\nspecified in argument and loads all of them before continuing. It is even\npossible to pass multiple "-f" arguments if the patterns are to be loaded from\nmultiple files. Empty lines as well as lines beginning with a sharp ('#') will\nbe ignored. All leading spaces and tabs will be stripped. If it is absolutely\nneeded to insert a valid pattern beginning with a sharp, just prefix it with a\nspace so that it is not taken for a comment. Depending on the data type and\nmatch method, haproxy may load the lines into a binary tree, allowing very fast\nlookups. This is true for IPv4 and exact string matching. In this case,\nduplicates will automatically be removed. Also, note that the "-i" flag applies\nto subsequent entries and not to entries loaded from files preceeding it. For\ninstance :\n\n acl valid-ua hdr(user-agent) -f exact-ua.lst -i -f generic-ua.lst test\n\nIn this example, each line of "exact-ua.lst" will be exactly matched against\nthe "user-agent" header of the request. Then each line of "generic-ua" will be\ncase-insensitively matched. Then the word "test" will be insensitively matched\ntoo.\n\nNote that right now it is difficult for the ACL parsers to report errors, so if\na file is unreadable or unparsable, the most you'll get is a parse error in the\nACL. Thus, file-based ACLs should only be produced by reliable processes.\n\nSupported types of values are :\n\n - integers or integer ranges\n - strings\n - regular expressions\n - IP addresses and networks\n\n\n7.1. Matching integers\n----------------------\n\nMatching integers is special in that ranges and operators are permitted. Note\nthat integer matching only applies to positive values. A range is a value\nexpressed with a lower and an upper bound separated with a colon, both of which\nmay be omitted.\n\nFor instance, "1024:65535" is a valid range to represent a range of\nunprivileged ports, and "1024:" would also work. "0:1023" is a valid\nrepresentation of privileged ports, and ":1023" would also work.\n\nAs a special case, some ACL functions support decimal numbers which are in fact\ntwo integers separated by a dot. This is used with some version checks for\ninstance. All integer properties apply to those decimal numbers, including\nranges and operators.\n\nFor an easier usage, comparison operators are also supported. Note that using\noperators with ranges does not make much sense and is strongly discouraged.\nSimilarly, it does not make much sense to perform order comparisons with a set\nof values.\n\nAvailable operators for integer matching are :\n\n eq : true if the tested value equals at least one value\n ge : true if the tested value is greater than or equal to at least one value\n gt : true if the tested value is greater than at least one value\n le : true if the tested value is less than or equal to at least one value\n lt : true if the tested value is less than at least one value\n\nFor instance, the following ACL matches any negative Content-Length header :\n\n acl negative-length hdr_val(content-length) lt 0\n\nThis one matches SSL versions between 3.0 and 3.1 (inclusive) :\n\n acl sslv3 req_ssl_ver 3:3.1\n\n\n7.2. Matching strings\n---------------------\n\nString matching applies to verbatim strings as they are passed, with the\nexception of the backslash ("\") which makes it possible to escape some\ncharacters such as the space. If the "-i" flag is passed before the first\nstring, then the matching will be performed ignoring the case. In order\nto match the string "-i", either set it second, or pass the "--" flag\nbefore the first string. Same applies of course to match the string "--".\n\n\n7.3. Matching regular expressions (regexes)\n-------------------------------------------\n\nJust like with string matching, regex matching applies to verbatim strings as\nthey are passed, with the exception of the backslash ("\") which makes it\npossible to escape some characters such as the space. If the "-i" flag is\npassed before the first regex, then the matching will be performed ignoring\nthe case. In order to match the string "-i", either set it second, or pass\nthe "--" flag before the first string. Same principle applies of course to\nmatch the string "--".\n\n\n7.4. Matching IPv4 addresses\n----------------------------\n\nIPv4 addresses values can be specified either as plain addresses or with a\nnetmask appended, in which case the IPv4 address matches whenever it is\nwithin the network. Plain addresses may also be replaced with a resolvable\nhost name, but this practice is generally discouraged as it makes it more\ndifficult to read and debug configurations. If hostnames are used, you should\nat least ensure that they are present in /etc/hosts so that the configuration\ndoes not depend on any random DNS match at the moment the configuration is\nparsed.\n\n\n7.5. Available matching criteria\n--------------------------------\n\n7.5.1. Matching at Layer 4 and below\n------------------------------------\n\nA first set of criteria applies to information which does not require any\nanalysis of the request or response contents. Those generally include TCP/IP\naddresses and ports, as well as internal values independant on the stream.\n\nalways_false\n This one never matches. All values and flags are ignored. It may be used as\n a temporary replacement for another one when adjusting configurations.\n\nalways_true\n This one always matches. All values and flags are ignored. It may be used as\n a temporary replacement for another one when adjusting configurations.\n\navg_queue <integer>\navg_queue(backend) <integer>\n Returns the total number of queued connections of the designated backend\n divided by the number of active servers. This is very similar to "queue"\n except that the size of the farm is considered, in order to give a more\n accurate measurement of the time it may take for a new connection to be\n processed. The main usage is to return a sorry page to new users when it\n becomes certain they will get a degraded service. Note that in the event\n there would not be any active server anymore, we would consider twice the\n number of queued connections as the measured value. This is a fair estimate,\n as we expect one server to get back soon anyway, but we still prefer to send\n new traffic to another backend if in better shape. See also the "queue",\n "be_conn", and "be_sess_rate" criteria.\n\nbe_conn <integer>\nbe_conn(backend) <integer>\n Applies to the number of currently established connections on the backend,\n possibly including the connection being evaluated. If no backend name is\n specified, the current one is used. But it is also possible to check another\n backend. It can be used to use a specific farm when the nominal one is full.\n See also the "fe_conn", "queue" and "be_sess_rate" criteria.\n\nbe_id <integer>\n Applies to the backend's id. Can be used in frontends to check from which\n backend it was called.\n\nbe_sess_rate <integer>\nbe_sess_rate(backend) <integer>\n Returns true when the sessions creation rate on the backend matches the\n specified values or ranges, in number of new sessions per second. This is\n used to switch to an alternate backend when an expensive or fragile one\n reaches too high a session rate, or to limit abuse of service (eg. prevent\n sucking of an online dictionary).\n\n Example :\n # Redirect to an error page if the dictionary is requested too often\n backend dynamic\n mode http\n acl being_scanned be_sess_rate gt 100\n redirect location /denied.html if being_scanned\n\nconnslots <integer>\nconnslots(backend) <integer>\n The basic idea here is to be able to measure the number of connection "slots"\n still available (connection + queue), so that anything beyond that (intended\n usage; see "use_backend" keyword) can be redirected to a different backend.\n\n 'connslots' = number of available server connection slots, + number of\n available server queue slots.\n\n Note that while "fe_conn" may be used, "connslots" comes in especially\n useful when you have a case of traffic going to one single ip, splitting into\n multiple backends (perhaps using acls to do name-based load balancing) and\n you want to be able to differentiate between different backends, and their\n available "connslots". Also, whereas "nbsrv" only measures servers that are\n actually *down*, this acl is more fine-grained and looks into the number of\n available connection slots as well. See also "queue" and "avg_queue".\n\n OTHER CAVEATS AND NOTES: at this point in time, the code does not take care\n of dynamic connections. Also, if any of the server maxconn, or maxqueue is 0,\n then this acl clearly does not make sense, in which case the value returned\n will be -1.\n\ndst <ip_address>\n Applies to the local IPv4 address the client connected to. It can be used to\n switch to a different backend for some alternative addresses.\n\ndst_conn <integer>\n Applies to the number of currently established connections on the same socket\n including the one being evaluated. It can be used to either return a sorry\n page before hard-blocking, or to use a specific backend to drain new requests\n when the socket is considered saturated. This offers the ability to assign\n different limits to different listening ports or addresses. See also the\n "fe_conn" and "be_conn" criteria.\n\ndst_port <integer>\n Applies to the local port the client connected to. It can be used to switch\n to a different backend for some alternative ports.\n\nfe_conn <integer>\nfe_conn(frontend) <integer>\n Applies to the number of currently established connections on the frontend,\n possibly including the connection being evaluated. If no frontend name is\n specified, the current one is used. But it is also possible to check another\n frontend. It can be used to either return a sorry page before hard-blocking,\n or to use a specific backend to drain new requests when the farm is\n considered saturated. See also the "dst_conn", "be_conn" and "fe_sess_rate"\n criteria.\n\nfe_id <integer>\n Applies to the frontend's id. Can be used in backends to check from which\n frontend it was called.\n\nfe_sess_rate <integer>\nfe_sess_rate(frontend) <integer>\n Returns true when the session creation rate on the current or the named\n frontend matches the specified values or ranges, expressed in new sessions\n per second. This is used to limit the connection rate to acceptable ranges in\n order to prevent abuse of service at the earliest moment. This can be\n combined with layer 4 ACLs in order to force the clients to wait a bit for\n the rate to go down below the limit.\n\n Example :\n # This frontend limits incoming mails to 10/s with a max of 100\n # concurrent connections. We accept any connection below 10/s, and\n # force excess clients to wait for 100 ms. Since clients are limited to\n # 100 max, there cannot be more than 10 incoming mails per second.\n frontend mail\n bind :25\n mode tcp\n maxconn 100\n acl too_fast fe_sess_rate ge 10\n tcp-request inspect-delay 100ms\n tcp-request content accept if ! too_fast\n tcp-request content accept if WAIT_END\n\nnbsrv <integer>\nnbsrv(backend) <integer>\n Returns true when the number of usable servers of either the current backend\n or the named backend matches the values or ranges specified. This is used to\n switch to an alternate backend when the number of servers is too low to\n to handle some load. It is useful to report a failure when combined with\n "monitor fail".\n\nqueue <integer>\nqueue(backend) <integer>\n Returns the total number of queued connections of the designated backend,\n including all the connections in server queues. If no backend name is\n specified, the current one is used, but it is also possible to check another\n one. This can be used to take actions when queuing goes above a known level,\n generally indicating a surge of traffic or a massive slowdown on the servers.\n One possible action could be to reject new users but still accept old ones.\n See also the "avg_queue", "be_conn", and "be_sess_rate" criteria.\n\nso_id <integer>\n Applies to the socket's id. Useful in frontends with many bind keywords.\n\nsrc <ip_address>\n Applies to the client's IPv4 address. It is usually used to limit access to\n certain resources such as statistics. Note that it is the TCP-level source\n address which is used, and not the address of a client behind a proxy.\n\nsrc_port <integer>\n Applies to the client's TCP source port. This has a very limited usage.\n\nsrv_id <integer>\n Applies to the server's id. Can be used in frontends or backends.\n\nsrv_is_up(<server>)\nsrv_is_up(<backend>/<server>)\n Returns true when the designated server is UP, and false when it is either\n DOWN or in maintenance mode. If <backend> is omitted, then the server is\n looked up in the current backend. The function takes no arguments since it\n is used as a boolean. It is mainly used to take action based on an external\n status reported via a health check (eg: a geographical site's availability).\n Another possible use which is more of a hack consists in using dummy servers\n as boolean variables that can be enabled or disabled from the CLI, so that\n rules depending on those ACLs can be tweaked in realtime.\n\n\n7.5.2. Matching contents at Layer 4\n-----------------------------------\n\nA second set of criteria depends on data found in buffers, but which can change\nduring analysis. This requires that some data has been buffered, for instance\nthrough TCP request content inspection. Please see the "tcp-request" keyword\nfor more detailed information on the subject.\n\nreq_len <integer>\n Returns true when the length of the data in the request buffer matches the\n specified range. It is important to understand that this test does not\n return false as long as the buffer is changing. This means that a check with\n equality to zero will almost always immediately match at the beginning of the\n session, while a test for more data will wait for that data to come in and\n return false only when haproxy is certain that no more data will come in.\n This test was designed to be used with TCP request content inspection.\n\nreq_proto_http\n Returns true when data in the request buffer look like HTTP and correctly\n parses as such. It is the same parser as the common HTTP request parser which\n is used so there should be no surprises. This test can be used for instance\n to direct HTTP traffic to a given port and HTTPS traffic to another one\n using TCP request content inspection rules.\n\nreq_rdp_cookie <string>\nreq_rdp_cookie(name) <string>\n Returns true when data in the request buffer look like the RDP protocol, and\n a cookie is present and equal to <string>. By default, any cookie name is\n checked, but a specific cookie name can be specified in parenthesis. The\n parser only checks for the first cookie, as illustrated in the RDP protocol\n specification. The cookie name is case insensitive. This ACL can be useful\n with the "MSTS" cookie, as it can contain the user name of the client\n connecting to the server if properly configured on the client. This can be\n used to restrict access to certain servers to certain users.\n\nreq_rdp_cookie_cnt <integer>\nreq_rdp_cookie_cnt(name) <integer>\n Returns true when the data in the request buffer look like the RDP protocol\n and the number of RDP cookies matches the specified range (typically zero or\n one). Optionally a specific cookie name can be checked. This is a simple way\n of detecting the RDP protocol, as clients generally send the MSTS or MSTSHASH\n cookies.\n\nreq_ssl_ver <decimal>\n Returns true when data in the request buffer look like SSL, with a protocol\n version matching the specified range. Both SSLv2 hello messages and SSLv3\n messages are supported. The test tries to be strict enough to avoid being\n easily fooled. In particular, it waits for as many bytes as announced in the\n message header if this header looks valid (bound to the buffer size). Note\n that TLSv1 is announced as SSL version 3.1. This test was designed to be used\n with TCP request content inspection.\n\nwait_end\n Waits for the end of the analysis period to return true. This may be used in\n conjunction with content analysis to avoid returning a wrong verdict early.\n It may also be used to delay some actions, such as a delayed reject for some\n special addresses. Since it either stops the rules evaluation or immediately\n returns true, it is recommended to use this acl as the last one in a rule.\n Please note that the default ACL "WAIT_END" is always usable without prior\n declaration. This test was designed to be used with TCP request content\n inspection.\n\n Examples :\n # delay every incoming request by 2 seconds\n tcp-request inspect-delay 2s\n tcp-request content accept if WAIT_END\n\n # don't immediately tell bad guys they are rejected\n tcp-request inspect-delay 10s\n acl goodguys src 10.0.0.0/24\n acl badguys src 10.0.1.0/24\n tcp-request content accept if goodguys\n tcp-request content reject if badguys WAIT_END\n tcp-request content reject\n\n\n7.5.3. Matching at Layer 7\n--------------------------\n\nA third set of criteria applies to information which can be found at the\napplication layer (layer 7). Those require that a full HTTP request has been\nread, and are only evaluated then. They may require slightly more CPU resources\nthan the layer 4 ones, but not much since the request and response are indexed.\n\nhdr <string>\nhdr(header) <string>\n Note: all the "hdr*" matching criteria either apply to all headers, or to a\n particular header whose name is passed between parenthesis and without any\n space. The header name is not case-sensitive. The header matching complies\n with RFC2616, and treats as separate headers all values delimited by commas.\n Use the shdr() variant for response headers sent by the server.\n\n The "hdr" criteria returns true if any of the headers matching the criteria\n match any of the strings. This can be used to check exact for values. For\n instance, checking that "connection: close" is set :\n\n hdr(Connection) -i close\n\nhdr_beg <string>\nhdr_beg(header) <string>\n Returns true when one of the headers begins with one of the strings. See\n "hdr" for more information on header matching. Use the shdr_beg() variant for\n response headers sent by the server.\n\nhdr_cnt <integer>\nhdr_cnt(header) <integer>\n Returns true when the number of occurrence of the specified header matches\n the values or ranges specified. It is important to remember that one header\n line may count as several headers if it has several values. This is used to\n detect presence, absence or abuse of a specific header, as well as to block\n request smuggling attacks by rejecting requests which contain more than one\n of certain headers. See "hdr" for more information on header matching. Use\n the shdr_cnt() variant for response headers sent by the server.\n\nhdr_dir <string>\nhdr_dir(header) <string>\n Returns true when one of the headers contains one of the strings either\n isolated or delimited by slashes. This is used to perform filename or\n directory name matching, and may be used with Referer. See "hdr" for more\n information on header matching. Use the shdr_dir() variant for response\n headers sent by the server.\n\nhdr_dom <string>\nhdr_dom(header) <string>\n Returns true when one of the headers contains one of the strings either\n isolated or delimited by dots. This is used to perform domain name matching,\n and may be used with the Host header. See "hdr" for more information on\n header matching. Use the shdr_dom() variant for response headers sent by the\n server.\n\nhdr_end <string>\nhdr_end(header) <string>\n Returns true when one of the headers ends with one of the strings. See "hdr"\n for more information on header matching. Use the shdr_end() variant for\n response headers sent by the server.\n\nhdr_ip <ip_address>\nhdr_ip(header) <ip_address>\n Returns true when one of the headers' values contains an IP address matching\n <ip_address>. This is mainly used with headers such as X-Forwarded-For or\n X-Client-IP. See "hdr" for more information on header matching. Use the\n shdr_ip() variant for response headers sent by the server.\n\nhdr_len <integer>\nhdr_len(<header>) <integer>\n Returns true when at least one of the headers has a length which matches the\n values or ranges specified. This may be used to detect empty or too large\n headers. See "hdr" for more information on header matching. Use the\n shdr_len() variant for response headers sent by the server.\n\nhdr_reg <regex>\nhdr_reg(header) <regex>\n Returns true when one of the headers matches of the regular expressions. It\n can be used at any time, but it is important to remember that regex matching\n is slower than other methods. See also other "hdr_" criteria, as well as\n "hdr" for more information on header matching. Use the shdr_reg() variant for\n response headers sent by the server.\n\nhdr_sub <string>\nhdr_sub(header) <string>\n Returns true when one of the headers contains one of the strings. See "hdr"\n for more information on header matching. Use the shdr_sub() variant for\n response headers sent by the server.\n\nhdr_val <integer>\nhdr_val(header) <integer>\n Returns true when one of the headers starts with a number which matches the\n values or ranges specified. This may be used to limit content-length to\n acceptable values for example. See "hdr" for more information on header\n matching. Use the shdr_val() variant for response headers sent by the server.\n\nhttp_auth(userlist)\nhttp_auth_group(userlist) <group> [<group>]*\n Returns true when authentication data received from the client matches\n username & password stored on the userlist. It is also possible to\n use http_auth_group to check if the user is assigned to at least one\n of specified groups.\n\n Currently only http basic auth is supported.\n\nhttp_first_req\n Returns true when the request being processed is the first one of the\n connection. This can be used to add or remove headers that may be missing\n from some requests when a request is not the first one, or even to perform\n some specific ACL checks only on the first request.\n\nmethod <string>\n Applies to the method in the HTTP request, eg: "GET". Some predefined ACL\n already check for most common methods.\n\npath <string>\n Returns true when the path part of the request, which starts at the first\n slash and ends before the question mark, equals one of the strings. It may be\n used to match known files, such as /favicon.ico.\n\npath_beg <string>\n Returns true when the path begins with one of the strings. This can be used\n to send certain directory names to alternative backends.\n\npath_dir <string>\n Returns true when one of the strings is found isolated or delimited with\n slashes in the path. This is used to perform filename or directory name\n matching without the risk of wrong match due to colliding prefixes. See also\n "url_dir" and "path_sub".\n\npath_dom <string>\n Returns true when one of the strings is found isolated or delimited with dots\n in the path. This may be used to perform domain name matching in proxy\n requests. See also "path_sub" and "url_dom".\n\npath_end <string>\n Returns true when the path ends with one of the strings. This may be used to\n control file name extension.\n\npath_len <integer>\n Returns true when the path length matches the values or ranges specified.\n This may be used to detect abusive requests for instance.\n\npath_reg <regex>\n Returns true when the path matches one of the regular expressions. It can be\n used any time, but it is important to remember that regex matching is slower\n than other methods. See also "url_reg" and all "path_" criteria.\n\npath_sub <string>\n Returns true when the path contains one of the strings. It can be used to\n detect particular patterns in paths, such as "../" for example. See also\n "path_dir".\n\nreq_ver <string>\n Applies to the version string in the HTTP request, eg: "1.0". Some predefined\n ACL already check for versions 1.0 and 1.1.\n\nstatus <integer>\n Applies to the HTTP status code in the HTTP response, eg: "302". It can be\n used to act on responses depending on status ranges, for instance, remove\n any Location header if the response is not a 3xx.\n\nurl <string>\n Applies to the whole URL passed in the request. The only real use is to match\n "*", for which there already is a predefined ACL.\n\nurl_beg <string>\n Returns true when the URL begins with one of the strings. This can be used to\n check whether a URL begins with a slash or with a protocol scheme.\n\nurl_dir <string>\n Returns true when one of the strings is found isolated or delimited with\n slashes in the URL. This is used to perform filename or directory name\n matching without the risk of wrong match due to colliding prefixes. See also\n "path_dir" and "url_sub".\n\nurl_dom <string>\n Returns true when one of the strings is found isolated or delimited with dots\n in the URL. This is used to perform domain name matching without the risk of\n wrong match due to colliding prefixes. See also "url_sub".\n\nurl_end <string>\n Returns true when the URL ends with one of the strings. It has very limited\n use. "path_end" should be used instead for filename matching.\n\nurl_ip <ip_address>\n Applies to the IP address specified in the absolute URI in an HTTP request.\n It can be used to prevent access to certain resources such as local network.\n It is useful with option "http_proxy".\n\nurl_len <integer>\n Returns true when the url length matches the values or ranges specified. This\n may be used to detect abusive requests for instance.\n\nurl_port <integer>\n Applies to the port specified in the absolute URI in an HTTP request. It can\n be used to prevent access to certain resources. It is useful with option\n "http_proxy". Note that if the port is not specified in the request, port 80\n is assumed.\n\nurl_reg <regex>\n Returns true when the URL matches one of the regular expressions. It can be\n used any time, but it is important to remember that regex matching is slower\n than other methods. See also "path_reg" and all "url_" criteria.\n\nurl_sub <string>\n Returns true when the URL contains one of the strings. It can be used to\n detect particular patterns in query strings for example. See also "path_sub".\n\n\n7.6. Pre-defined ACLs\n---------------------\n\nSome predefined ACLs are hard-coded so that they do not have to be declared in\nevery frontend which needs them. They all have their names in upper case in\norder to avoid confusion. Their equivalence is provided below.\n\nACL name Equivalent to Usage\n---------------+-----------------------------+---------------------------------\nFALSE always_false never match\nHTTP req_proto_http match if protocol is valid HTTP\nHTTP_1.0 req_ver 1.0 match HTTP version 1.0\nHTTP_1.1 req_ver 1.1 match HTTP version 1.1\nHTTP_CONTENT hdr_val(content-length) gt 0 match an existing content-length\nHTTP_URL_ABS url_reg ^[^/:]*:// match absolute URL with scheme\nHTTP_URL_SLASH url_beg / match URL beginning with "/"\nHTTP_URL_STAR url * match URL equal to "*"\nLOCALHOST src 127.0.0.1/8 match connection from local host\nMETH_CONNECT method CONNECT match HTTP CONNECT method\nMETH_GET method GET HEAD match HTTP GET or HEAD method\nMETH_HEAD method HEAD match HTTP HEAD method\nMETH_OPTIONS method OPTIONS match HTTP OPTIONS method\nMETH_POST method POST match HTTP POST method\nMETH_TRACE method TRACE match HTTP TRACE method\nRDP_COOKIE req_rdp_cookie_cnt gt 0 match presence of an RDP cookie\nREQ_CONTENT req_len gt 0 match data in the request buffer\nTRUE always_true always match\nWAIT_END wait_end wait for end of content analysis\n---------------+-----------------------------+---------------------------------\n\n\n7.7. Using ACLs to form conditions\n----------------------------------\n\nSome actions are only performed upon a valid condition. A condition is a\ncombination of ACLs with operators. 3 operators are supported :\n\n - AND (implicit)\n - OR (explicit with the "or" keyword or the "||" operator)\n - Negation with the exclamation mark ("!")\n\nA condition is formed as a disjunctive form:\n\n [!]acl1 [!]acl2 ... [!]acln { or [!]acl1 [!]acl2 ... [!]acln } ...\n\nSuch conditions are generally used after an "if" or "unless" statement,\nindicating when the condition will trigger the action.\n\nFor instance, to block HTTP requests to the "*" URL with methods other than\n"OPTIONS", as well as POST requests without content-length, and GET or HEAD\nrequests with a content-length greater than 0, and finally every request which\nis not either GET/HEAD/POST/OPTIONS !\n\n acl missing_cl hdr_cnt(Content-length) eq 0\n block if HTTP_URL_STAR !METH_OPTIONS || METH_POST missing_cl\n block if METH_GET HTTP_CONTENT\n block unless METH_GET or METH_POST or METH_OPTIONS\n\nTo select a different backend for requests to static contents on the "www" site\nand to every request on the "img", "video", "download" and "ftp" hosts :\n\n acl url_static path_beg /static /images /img /css\n acl url_static path_end .gif .png .jpg .css .js\n acl host_www hdr_beg(host) -i www\n acl host_static hdr_beg(host) -i img. video. download. ftp.\n\n # now use backend "static" for all static-only hosts, and for static urls\n # of host "www". Use backend "www" for the rest.\n use_backend static if host_static or host_www url_static\n use_backend www if host_www\n\nIt is also possible to form rules using "anonymous ACLs". Those are unnamed ACL\nexpressions that are built on the fly without needing to be declared. They must\nbe enclosed between braces, with a space before and after each brace (because\nthe braces must be seen as independant words). Example :\n\n The following rule :\n\n acl missing_cl hdr_cnt(Content-length) eq 0\n block if METH_POST missing_cl\n\n Can also be written that way :\n\n block if METH_POST { hdr_cnt(Content-length) eq 0 }\n\nIt is generally not recommended to use this construct because it's a lot easier\nto leave errors in the configuration when written that way. However, for very\nsimple rules matching only one source IP address for instance, it can make more\nsense to use them than to declare ACLs with random names. Another example of\ngood use is the following :\n\n With named ACLs :\n\n acl site_dead nbsrv(dynamic) lt 2\n acl site_dead nbsrv(static) lt 2\n monitor fail if site_dead\n\n With anonymous ACLs :\n\n monitor fail if { nbsrv(dynamic) lt 2 } || { nbsrv(static) lt 2 }\n\nSee section 4.2 for detailed help on the "block" and "use_backend" keywords.\n\n\n7.8. Pattern extraction\n-----------------------\n\nThe stickiness features relies on pattern extraction in the request and\nresponse. Sometimes the data needs to be converted first before being stored,\nfor instance converted from ASCII to IP or upper case to lower case.\n\nAll these operations of data extraction and conversion are defined as\n"pattern extraction rules". A pattern rule always has the same format. It\nbegins with a single pattern fetch word, potentially followed by a list of\narguments within parenthesis then an optional list of transformations. As\nmuch as possible, the pattern fetch functions use the same name as their\nequivalent used in ACLs.\n\nThe list of currently supported pattern fetch functions is the following :\n\n src This is the source IPv4 address of the client of the session.\n It is of type IP and only works with such tables.\n\n dst This is the destination IPv4 address of the session on the\n client side, which is the address the client connected to.\n It can be useful when running in transparent mode. It is of\n type IP and only works with such tables.\n\n dst_port This is the destination TCP port of the session on the client\n side, which is the port the client connected to. This might be\n used when running in transparent mode or when assigning dynamic\n ports to some clients for a whole application session. It is of\n type integer and only works with such tables.\n\n hdr(name) This extracts the last occurrence of header <name> in an HTTP\n request and converts it to an IP address. This IP address is\n then used to match the table. A typical use is with the\n x-forwarded-for header.\n\n\nThe currently available list of transformations include :\n\n lower Convert a string pattern to lower case. This can only be placed\n after a string pattern fetch function or after a conversion\n function returning a string type. The result is of type string.\n\n upper Convert a string pattern to upper case. This can only be placed\n after a string pattern fetch function or after a conversion\n function returning a string type. The result is of type string.\n\n ipmask(mask) Apply a mask to an IPv4 address, and use the result for lookups\n and storage. This can be used to make all hosts within a\n certain mask to share the same table entries and as such use\n the same server. The mask can be passed in dotted form (eg:\n 255.255.255.0) or in CIDR form (eg: 24).\n\n\n8. Logging\n----------\n\nOne of HAProxy's strong points certainly lies is its precise logs. It probably\nprovides the finest level of information available for such a product, which is\nvery important for troubleshooting complex environments. Standard information\nprovided in logs include client ports, TCP/HTTP state timers, precise session\nstate at termination and precise termination cause, information about decisions\nto direct traffic to a server, and of course the ability to capture arbitrary\nheaders.\n\nIn order to improve administrators reactivity, it offers a great transparency\nabout encountered problems, both internal and external, and it is possible to\nsend logs to different sources at the same time with different level filters :\n\n - global process-level logs (system errors, start/stop, etc..)\n - per-instance system and internal errors (lack of resource, bugs, ...)\n - per-instance external troubles (servers up/down, max connections)\n - per-instance activity (client connections), either at the establishment or\n at the termination.\n\nThe ability to distribute different levels of logs to different log servers\nallow several production teams to interact and to fix their problems as soon\nas possible. For example, the system team might monitor system-wide errors,\nwhile the application team might be monitoring the up/down for their servers in\nreal time, and the security team might analyze the activity logs with one hour\ndelay.\n\n\n8.1. Log levels\n---------------\n\nTCP and HTTP connections can be logged with information such as the date, time,\nsource IP address, destination address, connection duration, response times,\nHTTP request, HTTP return code, number of bytes transmitted, conditions\nin which the session ended, and even exchanged cookies values. For example\ntrack a particular user's problems. All messages may be sent to up to two\nsyslog servers. Check the "log" keyword in section 4.2 for more information\nabout log facilities.\n\n\n8.2. Log formats\n----------------\n\nHAProxy supports 4 log formats. Several fields are common between these formats\nand will be detailed in the following sections. A few of them may vary\nslightly with the configuration, due to indicators specific to certain\noptions. The supported formats are as follows :\n\n - the default format, which is very basic and very rarely used. It only\n provides very basic information about the incoming connection at the moment\n it is accepted : source IP:port, destination IP:port, and frontend-name.\n This mode will eventually disappear so it will not be described to great\n extents.\n\n - the TCP format, which is more advanced. This format is enabled when "option\n tcplog" is set on the frontend. HAProxy will then usually wait for the\n connection to terminate before logging. This format provides much richer\n information, such as timers, connection counts, queue size, etc... This\n format is recommended for pure TCP proxies.\n\n - the HTTP format, which is the most advanced for HTTP proxying. This format\n is enabled when "option httplog" is set on the frontend. It provides the\n same information as the TCP format with some HTTP-specific fields such as\n the request, the status code, and captures of headers and cookies. This\n format is recommended for HTTP proxies.\n\n - the CLF HTTP format, which is equivalent to the HTTP format, but with the\n fields arranged in the same order as the CLF format. In this mode, all\n timers, captures, flags, etc... appear one per field after the end of the\n common fields, in the same order they appear in the standard HTTP format.\n\nNext sections will go deeper into details for each of these formats. Format\nspecification will be performed on a "field" basis. Unless stated otherwise, a\nfield is a portion of text delimited by any number of spaces. Since syslog\nservers are susceptible of inserting fields at the beginning of a line, it is\nalways assumed that the first field is the one containing the process name and\nidentifier.\n\nNote : Since log lines may be quite long, the log examples in sections below\n might be broken into multiple lines. The example log lines will be\n prefixed with 3 closing angle brackets ('>>>') and each time a log is\n broken into multiple lines, each non-final line will end with a\n backslash ('\') and the next line will start indented by two characters.\n\n\n8.2.1. Default log format\n-------------------------\n\nThis format is used when no specific option is set. The log is emitted as soon\nas the connection is accepted. One should note that this currently is the only\nformat which logs the request's destination IP and ports.\n\n Example :\n listen www\n mode http\n log global\n server srv1 127.0.0.1:8000\n\n >>> Feb 6 12:12:09 localhost \\n haproxy[14385]: Connect from 10.0.1.2:33312 to 10.0.3.31:8012 \\n (www/HTTP)\n\n Field Format Extract from the example above\n 1 process_name '[' pid ']:' haproxy[14385]:\n 2 'Connect from' Connect from\n 3 source_ip ':' source_port 10.0.1.2:33312\n 4 'to' to\n 5 destination_ip ':' destination_port 10.0.3.31:8012\n 6 '(' frontend_name '/' mode ')' (www/HTTP)\n\nDetailed fields description :\n - "source_ip" is the IP address of the client which initiated the connection.\n - "source_port" is the TCP port of the client which initiated the connection.\n - "destination_ip" is the IP address the client connected to.\n - "destination_port" is the TCP port the client connected to.\n - "frontend_name" is the name of the frontend (or listener) which received\n and processed the connection.\n - "mode is the mode the frontend is operating (TCP or HTTP).\n\nIt is advised not to use this deprecated format for newer installations as it\nwill eventually disappear.\n\n\n8.2.2. TCP log format\n---------------------\n\nThe TCP format is used when "option tcplog" is specified in the frontend, and\nis the recommended format for pure TCP proxies. It provides a lot of precious\ninformation for troubleshooting. Since this format includes timers and byte\ncounts, the log is normally emitted at the end of the session. It can be\nemitted earlier if "option logasap" is specified, which makes sense in most\nenvironments with long sessions such as remote terminals. Sessions which match\nthe "monitor" rules are never logged. It is also possible not to emit logs for\nsessions for which no data were exchanged between the client and the server, by\nspecifying "option dontlognull" in the frontend. Successful connections will\nnot be logged if "option dontlog-normal" is specified in the frontend. A few\nfields may slightly vary depending on some configuration options, those are\nmarked with a star ('*') after the field name below.\n\n Example :\n frontend fnt\n mode tcp\n option tcplog\n log global\n default_backend bck\n\n backend bck\n server srv1 127.0.0.1:8000\n\n >>> Feb 6 12:12:56 localhost \\n haproxy[14387]: 10.0.1.2:33313 [06/Feb/2009:12:12:51.443] fnt \\n bck/srv1 0/0/5007 212 -- 0/0/0/0/3 0/0\n\n Field Format Extract from the example above\n 1 process_name '[' pid ']:' haproxy[14387]:\n 2 client_ip ':' client_port 10.0.1.2:33313\n 3 '[' accept_date ']' [06/Feb/2009:12:12:51.443]\n 4 frontend_name fnt\n 5 backend_name '/' server_name bck/srv1\n 6 Tw '/' Tc '/' Tt* 0/0/5007\n 7 bytes_read* 212\n 8 termination_state --\n 9 actconn '/' feconn '/' beconn '/' srv_conn '/' retries* 0/0/0/0/3\n 10 srv_queue '/' backend_queue 0/0\n\nDetailed fields description :\n - "client_ip" is the IP address of the client which initiated the TCP\n connection to haproxy.\n\n - "client_port" is the TCP port of the client which initiated the connection.\n\n - "accept_date" is the exact date when the connection was received by haproxy\n (which might be very slightly different from the date observed on the\n network if there was some queuing in the system's backlog). This is usually\n the same date which may appear in any upstream firewall's log.\n\n - "frontend_name" is the name of the frontend (or listener) which received\n and processed the connection.\n\n - "backend_name" is the name of the backend (or listener) which was selected\n to manage the connection to the server. This will be the same as the\n frontend if no switching rule has been applied, which is common for TCP\n applications.\n\n - "server_name" is the name of the last server to which the connection was\n sent, which might differ from the first one if there were connection errors\n and a redispatch occurred. Note that this server belongs to the backend\n which processed the request. If the connection was aborted before reaching\n a server, "<NOSRV>" is indicated instead of a server name.\n\n - "Tw" is the total time in milliseconds spent waiting in the various queues.\n It can be "-1" if the connection was aborted before reaching the queue.\n See "Timers" below for more details.\n\n - "Tc" is the total time in milliseconds spent waiting for the connection to\n establish to the final server, including retries. It can be "-1" if the\n connection was aborted before a connection could be established. See\n "Timers" below for more details.\n\n - "Tt" is the total time in milliseconds elapsed between the accept and the\n last close. It covers all possible processings. There is one exception, if\n "option logasap" was specified, then the time counting stops at the moment\n the log is emitted. In this case, a '+' sign is prepended before the value,\n indicating that the final one will be larger. See "Timers" below for more\n details.\n\n - "bytes_read" is the total number of bytes transmitted from the server to\n the client when the log is emitted. If "option logasap" is specified, the\n this value will be prefixed with a '+' sign indicating that the final one\n may be larger. Please note that this value is a 64-bit counter, so log\n analysis tools must be able to handle it without overflowing.\n\n - "termination_state" is the condition the session was in when the session\n ended. This indicates the session state, which side caused the end of\n session to happen, and for what reason (timeout, error, ...). The normal\n flags should be "--", indicating the session was closed by either end with\n no data remaining in buffers. See below "Session state at disconnection"\n for more details.\n\n - "actconn" is the total number of concurrent connections on the process when\n the session was logged. It it useful to detect when some per-process system\n limits have been reached. For instance, if actconn is close to 512 when\n multiple connection errors occur, chances are high that the system limits\n the process to use a maximum of 1024 file descriptors and that all of them\n are used. See section 3 "Global parameters" to find how to tune the system.\n\n - "feconn" is the total number of concurrent connections on the frontend when\n the session was logged. It is useful to estimate the amount of resource\n required to sustain high loads, and to detect when the frontend's "maxconn"\n has been reached. Most often when this value increases by huge jumps, it is\n because there is congestion on the backend servers, but sometimes it can be\n caused by a denial of service attack.\n\n - "beconn" is the total number of concurrent connections handled by the\n backend when the session was logged. It includes the total number of\n concurrent connections active on servers as well as the number of\n connections pending in queues. It is useful to estimate the amount of\n additional servers needed to support high loads for a given application.\n Most often when this value increases by huge jumps, it is because there is\n congestion on the backend servers, but sometimes it can be caused by a\n denial of service attack.\n\n - "srv_conn" is the total number of concurrent connections still active on\n the server when the session was logged. It can never exceed the server's\n configured "maxconn" parameter. If this value is very often close or equal\n to the server's "maxconn", it means that traffic regulation is involved a\n lot, meaning that either the server's maxconn value is too low, or that\n there aren't enough servers to process the load with an optimal response\n time. When only one of the server's "srv_conn" is high, it usually means\n that this server has some trouble causing the connections to take longer to\n be processed than on other servers.\n\n - "retries" is the number of connection retries experienced by this session\n when trying to connect to the server. It must normally be zero, unless a\n server is being stopped at the same moment the connection was attempted.\n Frequent retries generally indicate either a network problem between\n haproxy and the server, or a misconfigured system backlog on the server\n preventing new connections from being queued. This field may optionally be\n prefixed with a '+' sign, indicating that the session has experienced a\n redispatch after the maximal retry count has been reached on the initial\n server. In this case, the server name appearing in the log is the one the\n connection was redispatched to, and not the first one, though both may\n sometimes be the same in case of hashing for instance. So as a general rule\n of thumb, when a '+' is present in front of the retry count, this count\n should not be attributed to the logged server.\n\n - "srv_queue" is the total number of requests which were processed before\n this one in the server queue. It is zero when the request has not gone\n through the server queue. It makes it possible to estimate the approximate\n server's response time by dividing the time spent in queue by the number of\n requests in the queue. It is worth noting that if a session experiences a\n redispatch and passes through two server queues, their positions will be\n cumulated. A request should not pass through both the server queue and the\n backend queue unless a redispatch occurs.\n\n - "backend_queue" is the total number of requests which were processed before\n this one in the backend's global queue. It is zero when the request has not\n gone through the global queue. It makes it possible to estimate the average\n queue length, which easily translates into a number of missing servers when\n divided by a server's "maxconn" parameter. It is worth noting that if a\n session experiences a redispatch, it may pass twice in the backend's queue,\n and then both positions will be cumulated. A request should not pass\n through both the server queue and the backend queue unless a redispatch\n occurs.\n\n\n8.2.3. HTTP log format\n----------------------\n\nThe HTTP format is the most complete and the best suited for HTTP proxies. It\nis enabled by when "option httplog" is specified in the frontend. It provides\nthe same level of information as the TCP format with additional features which\nare specific to the HTTP protocol. Just like the TCP format, the log is usually\nemitted at the end of the session, unless "option logasap" is specified, which\ngenerally only makes sense for download sites. A session which matches the\n"monitor" rules will never logged. It is also possible not to log sessions for\nwhich no data were sent by the client by specifying "option dontlognull" in the\nfrontend. Successful connections will not be logged if "option dontlog-normal"\nis specified in the frontend.\n\nMost fields are shared with the TCP log, some being different. A few fields may\nslightly vary depending on some configuration options. Those ones are marked\nwith a star ('*') after the field name below.\n\n Example :\n frontend http-in\n mode http\n option httplog\n log global\n default_backend bck\n\n backend static\n server srv1 127.0.0.1:8000\n\n >>> Feb 6 12:14:14 localhost \\n haproxy[14389]: 10.0.1.2:33317 [06/Feb/2009:12:14:14.655] http-in \\n static/srv1 10/0/30/69/109 200 2750 - - ---- 1/1/1/1/0 0/0 {1wt.eu} \\n {} "GET /index.html HTTP/1.1"\n\n Field Format Extract from the example above\n 1 process_name '[' pid ']:' haproxy[14389]:\n 2 client_ip ':' client_port 10.0.1.2:33317\n 3 '[' accept_date ']' [06/Feb/2009:12:14:14.655]\n 4 frontend_name http-in\n 5 backend_name '/' server_name static/srv1\n 6 Tq '/' Tw '/' Tc '/' Tr '/' Tt* 10/0/30/69/109\n 7 status_code 200\n 8 bytes_read* 2750\n 9 captured_request_cookie -\n 10 captured_response_cookie -\n 11 termination_state ----\n 12 actconn '/' feconn '/' beconn '/' srv_conn '/' retries* 1/1/1/1/0\n 13 srv_queue '/' backend_queue 0/0\n 14 '{' captured_request_headers* '}' {haproxy.1wt.eu}\n 15 '{' captured_response_headers* '}' {}\n 16 '"' http_request '"' "GET /index.html HTTP/1.1"\n\n\nDetailed fields description :\n - "client_ip" is the IP address of the client which initiated the TCP\n connection to haproxy.\n\n - "client_port" is the TCP port of the client which initiated the connection.\n\n - "accept_date" is the exact date when the TCP connection was received by\n haproxy (which might be very slightly different from the date observed on\n the network if there was some queuing in the system's backlog). This is\n usually the same date which may appear in any upstream firewall's log. This\n does not depend on the fact that the client has sent the request or not.\n\n - "frontend_name" is the name of the frontend (or listener) which received\n and processed the connection.\n\n - "backend_name" is the name of the backend (or listener) which was selected\n to manage the connection to the server. This will be the same as the\n frontend if no switching rule has been applied.\n\n - "server_name" is the name of the last server to which the connection was\n sent, which might differ from the first one if there were connection errors\n and a redispatch occurred. Note that this server belongs to the backend\n which processed the request. If the request was aborted before reaching a\n server, "<NOSRV>" is indicated instead of a server name. If the request was\n intercepted by the stats subsystem, "<STATS>" is indicated instead.\n\n - "Tq" is the total time in milliseconds spent waiting for the client to send\n a full HTTP request, not counting data. It can be "-1" if the connection\n was aborted before a complete request could be received. It should always\n be very small because a request generally fits in one single packet. Large\n times here generally indicate network trouble between the client and\n haproxy. See "Timers" below for more details.\n\n - "Tw" is the total time in milliseconds spent waiting in the various queues.\n It can be "-1" if the connection was aborted before reaching the queue.\n See "Timers" below for more details.\n\n - "Tc" is the total time in milliseconds spent waiting for the connection to\n establish to the final server, including retries. It can be "-1" if the\n request was aborted before a connection could be established. See "Timers"\n below for more details.\n\n - "Tr" is the total time in milliseconds spent waiting for the server to send\n a full HTTP response, not counting data. It can be "-1" if the request was\n aborted before a complete response could be received. It generally matches\n the server's processing time for the request, though it may be altered by\n the amount of data sent by the client to the server. Large times here on\n "GET" requests generally indicate an overloaded server. See "Timers" below\n for more details.\n\n - "Tt" is the total time in milliseconds elapsed between the accept and the\n last close. It covers all possible processings. There is one exception, if\n "option logasap" was specified, then the time counting stops at the moment\n the log is emitted. In this case, a '+' sign is prepended before the value,\n indicating that the final one will be larger. See "Timers" below for more\n details.\n\n - "status_code" is the HTTP status code returned to the client. This status\n is generally set by the server, but it might also be set by haproxy when\n the server cannot be reached or when its response is blocked by haproxy.\n\n - "bytes_read" is the total number of bytes transmitted to the client when\n the log is emitted. This does include HTTP headers. If "option logasap" is\n specified, the this value will be prefixed with a '+' sign indicating that\n the final one may be larger. Please note that this value is a 64-bit\n counter, so log analysis tools must be able to handle it without\n overflowing.\n\n - "captured_request_cookie" is an optional "name=value" entry indicating that\n the client had this cookie in the request. The cookie name and its maximum\n length are defined by the "capture cookie" statement in the frontend\n configuration. The field is a single dash ('-') when the option is not\n set. Only one cookie may be captured, it is generally used to track session\n ID exchanges between a client and a server to detect session crossing\n between clients due to application bugs. For more details, please consult\n the section "Capturing HTTP headers and cookies" below.\n\n - "captured_response_cookie" is an optional "name=value" entry indicating\n that the server has returned a cookie with its response. The cookie name\n and its maximum length are defined by the "capture cookie" statement in the\n frontend configuration. The field is a single dash ('-') when the option is\n not set. Only one cookie may be captured, it is generally used to track\n session ID exchanges between a client and a server to detect session\n crossing between clients due to application bugs. For more details, please\n consult the section "Capturing HTTP headers and cookies" below.\n\n - "termination_state" is the condition the session was in when the session\n ended. This indicates the session state, which side caused the end of\n session to happen, for what reason (timeout, error, ...), just like in TCP\n logs, and information about persistence operations on cookies in the last\n two characters. The normal flags should begin with "--", indicating the\n session was closed by either end with no data remaining in buffers. See\n below "Session state at disconnection" for more details.\n\n - "actconn" is the total number of concurrent connections on the process when\n the session was logged. It it useful to detect when some per-process system\n limits have been reached. For instance, if actconn is close to 512 or 1024\n when multiple connection errors occur, chances are high that the system\n limits the process to use a maximum of 1024 file descriptors and that all\n of them are used. See section 3 "Global parameters" to find how to tune the\n system.\n\n - "feconn" is the total number of concurrent connections on the frontend when\n the session was logged. It is useful to estimate the amount of resource\n required to sustain high loads, and to detect when the frontend's "maxconn"\n has been reached. Most often when this value increases by huge jumps, it is\n because there is congestion on the backend servers, but sometimes it can be\n caused by a denial of service attack.\n\n - "beconn" is the total number of concurrent connections handled by the\n backend when the session was logged. It includes the total number of\n concurrent connections active on servers as well as the number of\n connections pending in queues. It is useful to estimate the amount of\n additional servers needed to support high loads for a given application.\n Most often when this value increases by huge jumps, it is because there is\n congestion on the backend servers, but sometimes it can be caused by a\n denial of service attack.\n\n - "srv_conn" is the total number of concurrent connections still active on\n the server when the session was logged. It can never exceed the server's\n configured "maxconn" parameter. If this value is very often close or equal\n to the server's "maxconn", it means that traffic regulation is involved a\n lot, meaning that either the server's maxconn value is too low, or that\n there aren't enough servers to process the load with an optimal response\n time. When only one of the server's "srv_conn" is high, it usually means\n that this server has some trouble causing the requests to take longer to be\n processed than on other servers.\n\n - "retries" is the number of connection retries experienced by this session\n when trying to connect to the server. It must normally be zero, unless a\n server is being stopped at the same moment the connection was attempted.\n Frequent retries generally indicate either a network problem between\n haproxy and the server, or a misconfigured system backlog on the server\n preventing new connections from being queued. This field may optionally be\n prefixed with a '+' sign, indicating that the session has experienced a\n redispatch after the maximal retry count has been reached on the initial\n server. In this case, the server name appearing in the log is the one the\n connection was redispatched to, and not the first one, though both may\n sometimes be the same in case of hashing for instance. So as a general rule\n of thumb, when a '+' is present in front of the retry count, this count\n should not be attributed to the logged server.\n\n - "srv_queue" is the total number of requests which were processed before\n this one in the server queue. It is zero when the request has not gone\n through the server queue. It makes it possible to estimate the approximate\n server's response time by dividing the time spent in queue by the number of\n requests in the queue. It is worth noting that if a session experiences a\n redispatch and passes through two server queues, their positions will be\n cumulated. A request should not pass through both the server queue and the\n backend queue unless a redispatch occurs.\n\n - "backend_queue" is the total number of requests which were processed before\n this one in the backend's global queue. It is zero when the request has not\n gone through the global queue. It makes it possible to estimate the average\n queue length, which easily translates into a number of missing servers when\n divided by a server's "maxconn" parameter. It is worth noting that if a\n session experiences a redispatch, it may pass twice in the backend's queue,\n and then both positions will be cumulated. A request should not pass\n through both the server queue and the backend queue unless a redispatch\n occurs.\n\n - "captured_request_headers" is a list of headers captured in the request due\n to the presence of the "capture request header" statement in the frontend.\n Multiple headers can be captured, they will be delimited by a vertical bar\n ('|'). When no capture is enabled, the braces do not appear, causing a\n shift of remaining fields. It is important to note that this field may\n contain spaces, and that using it requires a smarter log parser than when\n it's not used. Please consult the section "Capturing HTTP headers and\n cookies" below for more details.\n\n - "captured_response_headers" is a list of headers captured in the response\n due to the presence of the "capture response header" statement in the\n frontend. Multiple headers can be captured, they will be delimited by a\n vertical bar ('|'). When no capture is enabled, the braces do not appear,\n causing a shift of remaining fields. It is important to note that this\n field may contain spaces, and that using it requires a smarter log parser\n than when it's not used. Please consult the section "Capturing HTTP headers\n and cookies" below for more details.\n\n - "http_request" is the complete HTTP request line, including the method,\n request and HTTP version string. Non-printable characters are encoded (see\n below the section "Non-printable characters"). This is always the last\n field, and it is always delimited by quotes and is the only one which can\n contain quotes. If new fields are added to the log format, they will be\n added before this field. This field might be truncated if the request is\n huge and does not fit in the standard syslog buffer (1024 characters). This\n is the reason why this field must always remain the last one.\n\n\n8.3. Advanced logging options\n-----------------------------\n\nSome advanced logging options are often looked for but are not easy to find out\njust by looking at the various options. Here is an entry point for the few\noptions which can enable better logging. Please refer to the keywords reference\nfor more information about their usage.\n\n\n8.3.1. Disabling logging of external tests\n------------------------------------------\n\nIt is quite common to have some monitoring tools perform health checks on\nhaproxy. Sometimes it will be a layer 3 load-balancer such as LVS or any\ncommercial load-balancer, and sometimes it will simply be a more complete\nmonitoring system such as Nagios. When the tests are very frequent, users often\nask how to disable logging for those checks. There are three possibilities :\n\n - if connections come from everywhere and are just TCP probes, it is often\n desired to simply disable logging of connections without data exchange, by\n setting "option dontlognull" in the frontend. It also disables logging of\n port scans, which may or may not be desired.\n\n - if the connection come from a known source network, use "monitor-net" to\n declare this network as monitoring only. Any host in this network will then\n only be able to perform health checks, and their requests will not be\n logged. This is generally appropriate to designate a list of equipments\n such as other load-balancers.\n\n - if the tests are performed on a known URI, use "monitor-uri" to declare\n this URI as dedicated to monitoring. Any host sending this request will\n only get the result of a health-check, and the request will not be logged.\n\n\n8.3.2. Logging before waiting for the session to terminate\n----------------------------------------------------------\n\nThe problem with logging at end of connection is that you have no clue about\nwhat is happening during very long sessions, such as remote terminal sessions\nor large file downloads. This problem can be worked around by specifying\n"option logasap" in the frontend. Haproxy will then log as soon as possible,\njust before data transfer begins. This means that in case of TCP, it will still\nlog the connection status to the server, and in case of HTTP, it will log just\nafter processing the server headers. In this case, the number of bytes reported\nis the number of header bytes sent to the client. In order to avoid confusion\nwith normal logs, the total time field and the number of bytes are prefixed\nwith a '+' sign which means that real numbers are certainly larger.\n\n\n8.3.3. Raising log level upon errors\n------------------------------------\n\nSometimes it is more convenient to separate normal traffic from errors logs,\nfor instance in order to ease error monitoring from log files. When the option\n"log-separate-errors" is used, connections which experience errors, timeouts,\nretries, redispatches or HTTP status codes 5xx will see their syslog level\nraised from "info" to "err". This will help a syslog daemon store the log in\na separate file. It is very important to keep the errors in the normal traffic\nfile too, so that log ordering is not altered. You should also be careful if\nyou already have configured your syslog daemon to store all logs higher than\n"notice" in an "admin" file, because the "err" level is higher than "notice".\n\n\n8.3.4. Disabling logging of successful connections\n--------------------------------------------------\n\nAlthough this may sound strange at first, some large sites have to deal with\nmultiple thousands of logs per second and are experiencing difficulties keeping\nthem intact for a long time or detecting errors within them. If the option\n"dontlog-normal" is set on the frontend, all normal connections will not be\nlogged. In this regard, a normal connection is defined as one without any\nerror, timeout, retry nor redispatch. In HTTP, the status code is checked too,\nand a response with a status 5xx is not considered normal and will be logged\ntoo. Of course, doing is is really discouraged as it will remove most of the\nuseful information from the logs. Do this only if you have no other\nalternative.\n\n\n8.4. Timing events\n------------------\n\nTimers provide a great help in troubleshooting network problems. All values are\nreported in milliseconds (ms). These timers should be used in conjunction with\nthe session termination flags. In TCP mode with "option tcplog" set on the\nfrontend, 3 control points are reported under the form "Tw/Tc/Tt", and in HTTP\nmode, 5 control points are reported under the form "Tq/Tw/Tc/Tr/Tt" :\n\n - Tq: total time to get the client request (HTTP mode only). It's the time\n elapsed between the moment the client connection was accepted and the\n moment the proxy received the last HTTP header. The value "-1" indicates\n that the end of headers (empty line) has never been seen. This happens when\n the client closes prematurely or times out.\n\n - Tw: total time spent in the queues waiting for a connection slot. It\n accounts for backend queue as well as the server queues, and depends on the\n queue size, and the time needed for the server to complete previous\n requests. The value "-1" means that the request was killed before reaching\n the queue, which is generally what happens with invalid or denied requests.\n\n - Tc: total time to establish the TCP connection to the server. It's the time\n elapsed between the moment the proxy sent the connection request, and the\n moment it was acknowledged by the server, or between the TCP SYN packet and\n the matching SYN/ACK packet in return. The value "-1" means that the\n connection never established.\n\n - Tr: server response time (HTTP mode only). It's the time elapsed between\n the moment the TCP connection was established to the server and the moment\n the server sent its complete response headers. It purely shows its request\n processing time, without the network overhead due to the data transmission.\n It is worth noting that when the client has data to send to the server, for\n instance during a POST request, the time already runs, and this can distort\n apparent response time. For this reason, it's generally wise not to trust\n too much this field for POST requests initiated from clients behind an\n untrusted network. A value of "-1" here means that the last the response\n header (empty line) was never seen, most likely because the server timeout\n stroke before the server managed to process the request.\n\n - Tt: total session duration time, between the moment the proxy accepted it\n and the moment both ends were closed. The exception is when the "logasap"\n option is specified. In this case, it only equals (Tq+Tw+Tc+Tr), and is\n prefixed with a '+' sign. From this field, we can deduce "Td", the data\n transmission time, by substracting other timers when valid :\n\n Td = Tt - (Tq + Tw + Tc + Tr)\n\n Timers with "-1" values have to be excluded from this equation. In TCP\n mode, "Tq" and "Tr" have to be excluded too. Note that "Tt" can never be\n negative.\n\nThese timers provide precious indications on trouble causes. Since the TCP\nprotocol defines retransmit delays of 3, 6, 12... seconds, we know for sure\nthat timers close to multiples of 3s are nearly always related to lost packets\ndue to network problems (wires, negotiation, congestion). Moreover, if "Tt" is\nclose to a timeout value specified in the configuration, it often means that a\nsession has been aborted on timeout.\n\nMost common cases :\n\n - If "Tq" is close to 3000, a packet has probably been lost between the\n client and the proxy. This is very rare on local networks but might happen\n when clients are on far remote networks and send large requests. It may\n happen that values larger than usual appear here without any network cause.\n Sometimes, during an attack or just after a resource starvation has ended,\n haproxy may accept thousands of connections in a few milliseconds. The time\n spent accepting these connections will inevitably slightly delay processing\n of other connections, and it can happen that request times in the order of\n a few tens of milliseconds are measured after a few thousands of new\n connections have been accepted at once. Setting "option http-server-close"\n may display larger request times since "Tq" also measures the time spent\n waiting for additional requests.\n\n - If "Tc" is close to 3000, a packet has probably been lost between the\n server and the proxy during the server connection phase. This value should\n always be very low, such as 1 ms on local networks and less than a few tens\n of ms on remote networks.\n\n - If "Tr" is nearly always lower than 3000 except some rare values which seem\n to be the average majored by 3000, there are probably some packets lost\n between the proxy and the server.\n\n - If "Tt" is large even for small byte counts, it generally is because\n neither the client nor the server decides to close the connection, for\n instance because both have agreed on a keep-alive connection mode. In order\n to solve this issue, it will be needed to specify "option httpclose" on\n either the frontend or the backend. If the problem persists, it means that\n the server ignores the "close" connection mode and expects the client to\n close. Then it will be required to use "option forceclose". Having the\n smallest possible 'Tt' is important when connection regulation is used with\n the "maxconn" option on the servers, since no new connection will be sent\n to the server until another one is released.\n\nOther noticeable HTTP log cases ('xx' means any value to be ignored) :\n\n Tq/Tw/Tc/Tr/+Tt The "option logasap" is present on the frontend and the log\n was emitted before the data phase. All the timers are valid\n except "Tt" which is shorter than reality.\n\n -1/xx/xx/xx/Tt The client was not able to send a complete request in time\n or it aborted too early. Check the session termination flags\n then "timeout http-request" and "timeout client" settings.\n\n Tq/-1/xx/xx/Tt It was not possible to process the request, maybe because\n servers were out of order, because the request was invalid\n or forbidden by ACL rules. Check the session termination\n flags.\n\n Tq/Tw/-1/xx/Tt The connection could not establish on the server. Either it\n actively refused it or it timed out after Tt-(Tq+Tw) ms.\n Check the session termination flags, then check the\n "timeout connect" setting. Note that the tarpit action might\n return similar-looking patterns, with "Tw" equal to the time\n the client connection was maintained open.\n\n Tq/Tw/Tc/-1/Tt The server has accepted the connection but did not return\n a complete response in time, or it closed its connexion\n unexpectedly after Tt-(Tq+Tw+Tc) ms. Check the session\n termination flags, then check the "timeout server" setting.\n\n\n8.5. Session state at disconnection\n-----------------------------------\n\nTCP and HTTP logs provide a session termination indicator in the\n"termination_state" field, just before the number of active connections. It is\n2-characters long in TCP mode, and is extended to 4 characters in HTTP mode,\neach of which has a special meaning :\n\n - On the first character, a code reporting the first event which caused the\n session to terminate :\n\n C : the TCP session was unexpectedly aborted by the client.\n\n S : the TCP session was unexpectedly aborted by the server, or the\n server explicitly refused it.\n\n P : the session was prematurely aborted by the proxy, because of a\n connection limit enforcement, because a DENY filter was matched,\n because of a security check which detected and blocked a dangerous\n error in server response which might have caused information leak\n (eg: cacheable cookie), or because the response was processed by\n the proxy (redirect, stats, etc...).\n\n R : a resource on the proxy has been exhausted (memory, sockets, source\n ports, ...). Usually, this appears during the connection phase, and\n system logs should contain a copy of the precise error. If this\n happens, it must be considered as a very serious anomaly which\n should be fixed as soon as possible by any means.\n\n I : an internal error was identified by the proxy during a self-check.\n This should NEVER happen, and you are encouraged to report any log\n containing this, because this would almost certainly be a bug. It\n would be wise to preventively restart the process after such an\n event too, in case it would be caused by memory corruption.\n\n c : the client-side timeout expired while waiting for the client to\n send or receive data.\n\n s : the server-side timeout expired while waiting for the server to\n send or receive data.\n\n - : normal session completion, both the client and the server closed\n with nothing left in the buffers.\n\n - on the second character, the TCP or HTTP session state when it was closed :\n\n R : the proxy was waiting for a complete, valid REQUEST from the client\n (HTTP mode only). Nothing was sent to any server.\n\n Q : the proxy was waiting in the QUEUE for a connection slot. This can\n only happen when servers have a 'maxconn' parameter set. It can\n also happen in the global queue after a redispatch consecutive to\n a failed attempt to connect to a dying server. If no redispatch is\n reported, then no connection attempt was made to any server.\n\n C : the proxy was waiting for the CONNECTION to establish on the\n server. The server might at most have noticed a connection attempt.\n\n H : the proxy was waiting for complete, valid response HEADERS from the\n server (HTTP only).\n\n D : the session was in the DATA phase.\n\n L : the proxy was still transmitting LAST data to the client while the\n server had already finished. This one is very rare as it can only\n happen when the client dies while receiving the last packets.\n\n T : the request was tarpitted. It has been held open with the client\n during the whole "timeout tarpit" duration or until the client\n closed, both of which will be reported in the "Tw" timer.\n\n - : normal session completion after end of data transfer.\n\n - the third character tells whether the persistence cookie was provided by\n the client (only in HTTP mode) :\n\n N : the client provided NO cookie. This is usually the case for new\n visitors, so counting the number of occurrences of this flag in the\n logs generally indicate a valid trend for the site frequentation.\n\n I : the client provided an INVALID cookie matching no known server.\n This might be caused by a recent configuration change, mixed\n cookies between HTTP/HTTPS sites, persistence conditionally\n ignored, or an attack.\n\n D : the client provided a cookie designating a server which was DOWN,\n so either "option persist" was used and the client was sent to\n this server, or it was not set and the client was redispatched to\n another server.\n\n V : the client provided a VALID cookie, and was sent to the associated\n server.\n\n E : the client provided a valid cookie, but with a last date which was\n older than what is allowed by the "maxidle" cookie parameter, so\n the cookie is consider EXPIRED and is ignored. The request will be\n redispatched just as if there was no cookie.\n\n O : the client provided a valid cookie, but with a first date which was\n older than what is allowed by the "maxlife" cookie parameter, so\n the cookie is consider too OLD and is ignored. The request will be\n redispatched just as if there was no cookie.\n\n - : does not apply (no cookie set in configuration).\n\n - the last character reports what operations were performed on the persistence\n cookie returned by the server (only in HTTP mode) :\n\n N : NO cookie was provided by the server, and none was inserted either.\n\n I : no cookie was provided by the server, and the proxy INSERTED one.\n Note that in "cookie insert" mode, if the server provides a cookie,\n it will still be overwritten and reported as "I" here.\n\n U : the proxy UPDATED the last date in the cookie that was presented by\n the client. This can only happen in insert mode with "maxidle". It\n happens everytime there is activity at a different date than the\n date indicated in the cookie. If any other change happens, such as\n a redispatch, then the cookie will be marked as inserted instead.\n\n P : a cookie was PROVIDED by the server and transmitted as-is.\n\n R : the cookie provided by the server was REWRITTEN by the proxy, which\n happens in "cookie rewrite" or "cookie prefix" modes.\n\n D : the cookie provided by the server was DELETED by the proxy.\n\n - : does not apply (no cookie set in configuration).\n\nThe combination of the two first flags gives a lot of information about what\nwas happening when the session terminated, and why it did terminate. It can be\nhelpful to detect server saturation, network troubles, local system resource\nstarvation, attacks, etc...\n\nThe most common termination flags combinations are indicated below. They are\nalphabetically sorted, with the lowercase set just after the upper case for\neasier finding and understanding.\n\n Flags Reason\n\n -- Normal termination.\n\n CC The client aborted before the connection could be established to the\n server. This can happen when haproxy tries to connect to a recently\n dead (or unchecked) server, and the client aborts while haproxy is\n waiting for the server to respond or for "timeout connect" to expire.\n\n CD The client unexpectedly aborted during data transfer. This can be\n caused by a browser crash, by an intermediate equipment between the\n client and haproxy which decided to actively break the connection,\n by network routing issues between the client and haproxy, or by a\n keep-alive session between the server and the client terminated first\n by the client.\n\n cD The client did not send nor acknowledge any data for as long as the\n "timeout client" delay. This is often caused by network failures on\n the client side, or the client simply leaving the net uncleanly.\n\n CH The client aborted while waiting for the server to start responding.\n It might be the server taking too long to respond or the client\n clicking the 'Stop' button too fast.\n\n cH The "timeout client" stroke while waiting for client data during a\n POST request. This is sometimes caused by too large TCP MSS values\n for PPPoE networks which cannot transport full-sized packets. It can\n also happen when client timeout is smaller than server timeout and\n the server takes too long to respond.\n\n CQ The client aborted while its session was queued, waiting for a server\n with enough empty slots to accept it. It might be that either all the\n servers were saturated or that the assigned server was taking too\n long a time to respond.\n\n CR The client aborted before sending a full HTTP request. Most likely\n the request was typed by hand using a telnet client, and aborted\n too early. The HTTP status code is likely a 400 here. Sometimes this\n might also be caused by an IDS killing the connection between haproxy\n and the client.\n\n cR The "timeout http-request" stroke before the client sent a full HTTP\n request. This is sometimes caused by too large TCP MSS values on the\n client side for PPPoE networks which cannot transport full-sized\n packets, or by clients sending requests by hand and not typing fast\n enough, or forgetting to enter the empty line at the end of the\n request. The HTTP status code is likely a 408 here.\n\n CT The client aborted while its session was tarpitted. It is important to\n check if this happens on valid requests, in order to be sure that no\n wrong tarpit rules have been written. If a lot of them happen, it\n might make sense to lower the "timeout tarpit" value to something\n closer to the average reported "Tw" timer, in order not to consume\n resources for just a few attackers.\n\n SC The server or an equipment between it and haproxy explicitly refused\n the TCP connection (the proxy received a TCP RST or an ICMP message\n in return). Under some circumstances, it can also be the network\n stack telling the proxy that the server is unreachable (eg: no route,\n or no ARP response on local network). When this happens in HTTP mode,\n the status code is likely a 502 or 503 here.\n\n sC The "timeout connect" stroke before a connection to the server could\n complete. When this happens in HTTP mode, the status code is likely a\n 503 or 504 here.\n\n SD The connection to the server died with an error during the data\n transfer. This usually means that haproxy has received an RST from\n the server or an ICMP message from an intermediate equipment while\n exchanging data with the server. This can be caused by a server crash\n or by a network issue on an intermediate equipment.\n\n sD The server did not send nor acknowledge any data for as long as the\n "timeout server" setting during the data phase. This is often caused\n by too short timeouts on L4 equipments before the server (firewalls,\n load-balancers, ...), as well as keep-alive sessions maintained\n between the client and the server expiring first on haproxy.\n\n SH The server aborted before sending its full HTTP response headers, or\n it crashed while processing the request. Since a server aborting at\n this moment is very rare, it would be wise to inspect its logs to\n control whether it crashed and why. The logged request may indicate a\n small set of faulty requests, demonstrating bugs in the application.\n Sometimes this might also be caused by an IDS killing the connection\n between haproxy and the server.\n\n sH The "timeout server" stroke before the server could return its\n response headers. This is the most common anomaly, indicating too\n long transactions, probably caused by server or database saturation.\n The immediate workaround consists in increasing the "timeout server"\n setting, but it is important to keep in mind that the user experience\n will suffer from these long response times. The only long term\n solution is to fix the application.\n\n sQ The session spent too much time in queue and has been expired. See\n the "timeout queue" and "timeout connect" settings to find out how to\n fix this if it happens too often. If it often happens massively in\n short periods, it may indicate general problems on the affected\n servers due to I/O or database congestion, or saturation caused by\n external attacks.\n\n PC The proxy refused to establish a connection to the server because the\n process' socket limit has been reached while attempting to connect.\n The global "maxconn" parameter may be increased in the configuration\n so that it does not happen anymore. This status is very rare and\n might happen when the global "ulimit-n" parameter is forced by hand.\n\n PD The proxy blocked an incorrectly formatted chunked encoded message in\n a request or a response, after the server has emitted its headers. In\n most cases, this will indicate an invalid message from the server to\n the client.\n\n PH The proxy blocked the server's response, because it was invalid,\n incomplete, dangerous (cache control), or matched a security filter.\n In any case, an HTTP 502 error is sent to the client. One possible\n cause for this error is an invalid syntax in an HTTP header name\n containing unauthorized characters. It is also possible but quite\n rare, that the proxy blocked a chunked-encoding request from the\n client due to an invalid syntax, before the server responded. In this\n case, an HTTP 400 error is sent to the client and reported in the\n logs.\n\n PR The proxy blocked the client's HTTP request, either because of an\n invalid HTTP syntax, in which case it returned an HTTP 400 error to\n the client, or because a deny filter matched, in which case it\n returned an HTTP 403 error.\n\n PT The proxy blocked the client's request and has tarpitted its\n connection before returning it a 500 server error. Nothing was sent\n to the server. The connection was maintained open for as long as\n reported by the "Tw" timer field.\n\n RC A local resource has been exhausted (memory, sockets, source ports)\n preventing the connection to the server from establishing. The error\n logs will tell precisely what was missing. This is very rare and can\n only be solved by proper system tuning.\n\nThe combination of the two last flags gives a lot of information about how\npersistence was handled by the client, the server and by haproxy. This is very\nimportant to troubleshoot disconnections, when users complain they have to\nre-authenticate. The commonly encountered flags are :\n\n -- Persistence cookie is not enabled.\n\n NN No cookie was provided by the client, none was inserted in the\n response. For instance, this can be in insert mode with "postonly"\n set on a GET request.\n\n II A cookie designating an invalid server was provided by the client,\n a valid one was inserted in the response. This typically happens when\n a "server" entry is removed from the configuraton, since its cookie\n value can be presented by a client when no other server knows it.\n\n NI No cookie was provided by the client, one was inserted in the\n response. This typically happens for first requests from every user\n in "insert" mode, which makes it an easy way to count real users.\n\n VN A cookie was provided by the client, none was inserted in the\n response. This happens for most responses for which the client has\n already got a cookie.\n\n VU A cookie was provided by the client, with a last visit date which is\n not completely up-to-date, so an updated cookie was provided in\n response. This can also happen if there was no date at all, or if\n there was a date but the "maxidle" parameter was not set, so that the\n cookie can be switched to unlimited time.\n\n EI A cookie was provided by the client, with a last visit date which is\n too old for the "maxidle" parameter, so the cookie was ignored and a\n new cookie was inserted in the response.\n\n OI A cookie was provided by the client, with a first visit date which is\n too old for the "maxlife" parameter, so the cookie was ignored and a\n new cookie was inserted in the response.\n\n DI The server designated by the cookie was down, a new server was\n selected and a new cookie was emitted in the response.\n\n VI The server designated by the cookie was not marked dead but could not\n be reached. A redispatch happened and selected another one, which was\n then advertised in the response.\n\n\n8.6. Non-printable characters\n-----------------------------\n\nIn order not to cause trouble to log analysis tools or terminals during log\nconsulting, non-printable characters are not sent as-is into log files, but are\nconverted to the two-digits hexadecimal representation of their ASCII code,\nprefixed by the character '#'. The only characters that can be logged without\nbeing escaped are comprised between 32 and 126 (inclusive). Obviously, the\nescape character '#' itself is also encoded to avoid any ambiguity ("#23"). It\nis the same for the character '"' which becomes "#22", as well as '{', '|' and\n'}' when logging headers.\n\nNote that the space character (' ') is not encoded in headers, which can cause\nissues for tools relying on space count to locate fields. A typical header\ncontaining spaces is "User-Agent".\n\nLast, it has been observed that some syslog daemons such as syslog-ng escape\nthe quote ('"') with a backslash ('\'). The reverse operation can safely be\nperformed since no quote may appear anywhere else in the logs.\n\n\n8.7. Capturing HTTP cookies\n---------------------------\n\nCookie capture simplifies the tracking a complete user session. This can be\nachieved using the "capture cookie" statement in the frontend. Please refer to\nsection 4.2 for more details. Only one cookie can be captured, and the same\ncookie will simultaneously be checked in the request ("Cookie:" header) and in\nthe response ("Set-Cookie:" header). The respective values will be reported in\nthe HTTP logs at the "captured_request_cookie" and "captured_response_cookie"\nlocations (see section 8.2.3 about HTTP log format). When either cookie is\nnot seen, a dash ('-') replaces the value. This way, it's easy to detect when a\nuser switches to a new session for example, because the server will reassign it\na new cookie. It is also possible to detect if a server unexpectedly sets a\nwrong cookie to a client, leading to session crossing.\n\n Examples :\n # capture the first cookie whose name starts with "ASPSESSION"\n capture cookie ASPSESSION len 32\n\n # capture the first cookie whose name is exactly "vgnvisitor"\n capture cookie vgnvisitor= len 32\n\n\n8.8. Capturing HTTP headers\n---------------------------\n\nHeader captures are useful to track unique request identifiers set by an upper\nproxy, virtual host names, user-agents, POST content-length, referrers, etc. In\nthe response, one can search for information about the response length, how the\nserver asked the cache to behave, or an object location during a redirection.\n\nHeader captures are performed using the "capture request header" and "capture\nresponse header" statements in the frontend. Please consult their definition in\nsection 4.2 for more details.\n\nIt is possible to include both request headers and response headers at the same\ntime. Non-existent headers are logged as empty strings, and if one header\nappears more than once, only its last occurrence will be logged. Request headers\nare grouped within braces '{' and '}' in the same order as they were declared,\nand delimited with a vertical bar '|' without any space. Response headers\nfollow the same representation, but are displayed after a space following the\nrequest headers block. These blocks are displayed just before the HTTP request\nin the logs.\n\n Example :\n # This instance chains to the outgoing proxy\n listen proxy-out\n mode http\n option httplog\n option logasap\n log global\n server cache1 192.168.1.1:3128\n\n # log the name of the virtual server\n capture request header Host len 20\n\n # log the amount of data uploaded during a POST\n capture request header Content-Length len 10\n\n # log the beginning of the referrer\n capture request header Referer len 20\n\n # server name (useful for outgoing proxies only)\n capture response header Server len 20\n\n # logging the content-length is useful with "option logasap"\n capture response header Content-Length len 10\n\n # log the expected cache behaviour on the response\n capture response header Cache-Control len 8\n\n # the Via header will report the next proxy's name\n capture response header Via len 20\n\n # log the URL location during a redirection\n capture response header Location len 20\n\n >>> Aug 9 20:26:09 localhost \\n haproxy[2022]: 127.0.0.1:34014 [09/Aug/2004:20:26:09] proxy-out \\n proxy-out/cache1 0/0/0/162/+162 200 +350 - - ---- 0/0/0/0/0 0/0 \\n {fr.adserver.yahoo.co||http://fr.f416.mail.} {|864|private||} \\n "GET http://fr.adserver.yahoo.com/"\n\n >>> Aug 9 20:30:46 localhost \\n haproxy[2022]: 127.0.0.1:34020 [09/Aug/2004:20:30:46] proxy-out \\n proxy-out/cache1 0/0/0/182/+182 200 +279 - - ---- 0/0/0/0/0 0/0 \\n {w.ods.org||} {Formilux/0.1.8|3495|||} \\n "GET http://trafic.1wt.eu/ HTTP/1.1"\n\n >>> Aug 9 20:30:46 localhost \\n haproxy[2022]: 127.0.0.1:34028 [09/Aug/2004:20:30:46] proxy-out \\n proxy-out/cache1 0/0/2/126/+128 301 +223 - - ---- 0/0/0/0/0 0/0 \\n {www.sytadin.equipement.gouv.fr||http://trafic.1wt.eu/} \\n {Apache|230|||http://www.sytadin.} \\n "GET http://www.sytadin.equipement.gouv.fr/ HTTP/1.1"\n\n\n8.9. Examples of logs\n---------------------\n\nThese are real-world examples of logs accompanied with an explanation. Some of\nthem have been made up by hand. The syslog part has been removed for better\nreading. Their sole purpose is to explain how to decipher them.\n\n >>> haproxy[674]: 127.0.0.1:33318 [15/Oct/2003:08:31:57.130] px-http \\n px-http/srv1 6559/0/7/147/6723 200 243 - - ---- 5/3/3/1/0 0/0 \\n "HEAD / HTTP/1.0"\n\n => long request (6.5s) entered by hand through 'telnet'. The server replied\n in 147 ms, and the session ended normally ('----')\n\n >>> haproxy[674]: 127.0.0.1:33319 [15/Oct/2003:08:31:57.149] px-http \\n px-http/srv1 6559/1230/7/147/6870 200 243 - - ---- 324/239/239/99/0 \\n 0/9 "HEAD / HTTP/1.0"\n\n => Idem, but the request was queued in the global queue behind 9 other\n requests, and waited there for 1230 ms.\n\n >>> haproxy[674]: 127.0.0.1:33320 [15/Oct/2003:08:32:17.654] px-http \\n px-http/srv1 9/0/7/14/+30 200 +243 - - ---- 3/3/3/1/0 0/0 \\n "GET /image.iso HTTP/1.0"\n\n => request for a long data transfer. The "logasap" option was specified, so\n the log was produced just before transferring data. The server replied in\n 14 ms, 243 bytes of headers were sent to the client, and total time from\n accept to first data byte is 30 ms.\n\n >>> haproxy[674]: 127.0.0.1:33320 [15/Oct/2003:08:32:17.925] px-http \\n px-http/srv1 9/0/7/14/30 502 243 - - PH-- 3/2/2/0/0 0/0 \\n "GET /cgi-bin/bug.cgi? HTTP/1.0"\n\n => the proxy blocked a server response either because of an "rspdeny" or\n "rspideny" filter, or because the response was improperly formatted and\n not HTTP-compliant, or because it blocked sensitive information which\n risked being cached. In this case, the response is replaced with a "502\n bad gateway". The flags ("PH--") tell us that it was haproxy who decided\n to return the 502 and not the server.\n\n >>> haproxy[18113]: 127.0.0.1:34548 [15/Oct/2003:15:18:55.798] px-http \\n px-http/<NOSRV> -1/-1/-1/-1/8490 -1 0 - - CR-- 2/2/2/0/0 0/0 ""\n\n => the client never completed its request and aborted itself ("C---") after\n 8.5s, while the proxy was waiting for the request headers ("-R--").\n Nothing was sent to any server.\n\n >>> haproxy[18113]: 127.0.0.1:34549 [15/Oct/2003:15:19:06.103] px-http \\n px-http/<NOSRV> -1/-1/-1/-1/50001 408 0 - - cR-- 2/2/2/0/0 0/0 ""\n\n => The client never completed its request, which was aborted by the\n time-out ("c---") after 50s, while the proxy was waiting for the request\n headers ("-R--"). Nothing was sent to any server, but the proxy could\n send a 408 return code to the client.\n\n >>> haproxy[18989]: 127.0.0.1:34550 [15/Oct/2003:15:24:28.312] px-tcp \\n px-tcp/srv1 0/0/5007 0 cD 0/0/0/0/0 0/0\n\n => This log was produced with "option tcplog". The client timed out after\n 5 seconds ("c----").\n\n >>> haproxy[18989]: 10.0.0.1:34552 [15/Oct/2003:15:26:31.462] px-http \\n px-http/srv1 3183/-1/-1/-1/11215 503 0 - - SC-- 205/202/202/115/3 \\n 0/0 "HEAD / HTTP/1.0"\n\n => The request took 3s to complete (probably a network problem), and the\n connection to the server failed ('SC--') after 4 attempts of 2 seconds\n (config says 'retries 3'), and no redispatch (otherwise we would have\n seen "/+3"). Status code 503 was returned to the client. There were 115\n connections on this server, 202 connections on this proxy, and 205 on\n the global process. It is possible that the server refused the\n connection because of too many already established.\n\n\n9. Statistics and monitoring\n----------------------------\n\nIt is possible to query HAProxy about its status. The most commonly used\nmechanism is the HTTP statistics page. This page also exposes an alternative\nCSV output format for monitoring tools. The same format is provided on the\nUnix socket.\n\n\n9.1. CSV format\n---------------\n\nThe statistics may be consulted either from the unix socket or from the HTTP\npage. Both means provide a CSV format whose fields follow.\n\n 0. pxname: proxy name\n 1. svname: service name (FRONTEND for frontend, BACKEND for backend, any name\n for server)\n 2. qcur: current queued requests\n 3. qmax: max queued requests\n 4. scur: current sessions\n 5. smax: max sessions\n 6. slim: sessions limit\n 7. stot: total sessions\n 8. bin: bytes in\n 9. bout: bytes out\n 10. dreq: denied requests\n 11. dresp: denied responses\n 12. ereq: request errors\n 13. econ: connection errors\n 14. eresp: response errors (among which srv_abrt)\n 15. wretr: retries (warning)\n 16. wredis: redispatches (warning)\n 17. status: status (UP/DOWN/NOLB/MAINT/MAINT(via)...)\n 18. weight: server weight (server), total weight (backend)\n 19. act: server is active (server), number of active servers (backend)\n 20. bck: server is backup (server), number of backup servers (backend)\n 21. chkfail: number of failed checks\n 22. chkdown: number of UP->DOWN transitions\n 23. lastchg: last status change (in seconds)\n 24. downtime: total downtime (in seconds)\n 25. qlimit: queue limit\n 26. pid: process id (0 for first instance, 1 for second, ...)\n 27. iid: unique proxy id\n 28. sid: service id (unique inside a proxy)\n 29. throttle: warm up status\n 30. lbtot: total number of times a server was selected\n 31. tracked: id of proxy/server if tracking is enabled\n 32. type (0=frontend, 1=backend, 2=server, 3=socket)\n 33. rate: number of sessions per second over last elapsed second\n 34. rate_lim: limit on new sessions per second\n 35. rate_max: max number of new sessions per second\n 36. check_status: status of last health check, one of:\n UNK -> unknown\n INI -> initializing\n SOCKERR -> socket error\n L4OK -> check passed on layer 4, no upper layers testing enabled\n L4TMOUT -> layer 1-4 timeout\n L4CON -> layer 1-4 connection problem, for example\n "Connection refused" (tcp rst) or "No route to host" (icmp)\n L6OK -> check passed on layer 6\n L6TOUT -> layer 6 (SSL) timeout\n L6RSP -> layer 6 invalid response - protocol error\n L7OK -> check passed on layer 7\n L7OKC -> check conditionally passed on layer 7, for example 404 with\n disable-on-404\n L7TOUT -> layer 7 (HTTP/SMTP) timeout\n L7RSP -> layer 7 invalid response - protocol error\n L7STS -> layer 7 response error, for example HTTP 5xx\n 37. check_code: layer5-7 code, if available\n 38. check_duration: time in ms took to finish last health check\n 39. hrsp_1xx: http responses with 1xx code\n 40. hrsp_2xx: http responses with 2xx code\n 41. hrsp_3xx: http responses with 3xx code\n 42. hrsp_4xx: http responses with 4xx code\n 43. hrsp_5xx: http responses with 5xx code\n 44. hrsp_other: http responses with other codes (protocol error)\n 45. hanafail: failed health checks details\n 46. req_rate: HTTP requests per second over last elapsed second\n 47. req_rate_max: max number of HTTP requests per second observed\n 48. req_tot: total number of HTTP requests received\n 49. cli_abrt: number of data transfers aborted by the client\n 50. srv_abrt: number of data transfers aborted by the server (inc. in eresp)\n\n\n9.2. Unix Socket commands\n-------------------------\n\nThe following commands are supported on the UNIX stats socket ; all of them\nmust be terminated by a line feed. The socket supports pipelining, so that it\nis possible to chain multiple commands at once provided they are delimited by\na semi-colon or a line feed, although the former is more reliable as it has no\nrisk of being truncated over the network. The responses themselves will each be\nfollowed by an empty line, so it will be easy for an external script to match a\ngiven response with a given request. By default one command line is processed\nthen the connection closes, but there is an interactive allowing multiple lines\nto be issued one at a time.\n\nIt is important to understand that when multiple haproxy processes are started\non the same sockets, any process may pick up the request and will output its\nown stats.\n\nclear counters\n Clear the max values of the statistics counters in each proxy (frontend &\n backend) and in each server. The cumulated counters are not affected. This\n can be used to get clean counters after an incident, without having to\n restart nor to clear traffic counters. This command is restricted and can\n only be issued on sockets configured for levels "operator" or "admin".\n\nclear counters all\n Clear all statistics counters in each proxy (frontend & backend) and in each\n server. This has the same effect as restarting. This command is restricted\n and can only be issued on sockets configured for level "admin".\n\ndisable server <backend>/<server>\n Mark the server DOWN for maintenance. In this mode, no more checks will be\n performed on the server until it leaves maintenance.\n If the server is tracked by other servers, those servers will be set to DOWN\n during the maintenance.\n\n In the statistics page, a server DOWN for maintenance will appear with a\n "MAINT" status, its tracking servers with the "MAINT(via)" one.\n\n Both the backend and the server may be specified either by their name or by\n their numeric ID, prefixed with a sharp ('#').\n\n This command is restricted and can only be issued on sockets configured for\n level "admin".\n\nenable server <backend>/<server>\n If the server was previously marked as DOWN for maintenance, this marks the\n server UP and checks are re-enabled.\n\n Both the backend and the server may be specified either by their name or by\n their numeric ID, prefixed with a sharp ('#').\n\n This command is restricted and can only be issued on sockets configured for\n level "admin".\n\nget weight <backend>/<server>\n Report the current weight and the initial weight of server <server> in\n backend <backend> or an error if either doesn't exist. The initial weight is\n the one that appears in the configuration file. Both are normally equal\n unless the current weight has been changed. Both the backend and the server\n may be specified either by their name or by their numeric ID, prefixed with a\n sharp ('#').\n\nhelp\n Print the list of known keywords and their basic usage. The same help screen\n is also displayed for unknown commands.\n\nprompt\n Toggle the prompt at the beginning of the line and enter or leave interactive\n mode. In interactive mode, the connection is not closed after a command\n completes. Instead, the prompt will appear again, indicating the user that\n the interpreter is waiting for a new command. The prompt consists in a right\n angle bracket followed by a space "> ". This mode is particularly convenient\n when one wants to periodically check information such as stats or errors.\n It is also a good idea to enter interactive mode before issuing a "help"\n command.\n\nquit\n Close the connection when in interactive mode.\n\nset timeout cli <delay>\n Change the CLI interface timeout for current connection. This can be useful\n during long debugging sessions where the user needs to constantly inspect\n some indicators without being disconnected. The delay is passed in seconds.\n\nset weight <backend>/<server> <weight>[%]\n Change a server's weight to the value passed in argument. If the value ends\n with the '%' sign, then the new weight will be relative to the initially\n configured weight. Relative weights are only permitted between 0 and 100%,\n and absolute weights are permitted between 0 and 256. Servers which are part\n of a farm running a static load-balancing algorithm have stricter limitations\n because the weight cannot change once set. Thus for these servers, the only\n accepted values are 0 and 100% (or 0 and the initial weight). Changes take\n effect immediately, though certain LB algorithms require a certain amount of\n requests to consider changes. A typical usage of this command is to disable\n a server during an update by setting its weight to zero, then to enable it\n again after the update by setting it back to 100%. This command is restricted\n and can only be issued on sockets configured for level "admin". Both the\n backend and the server may be specified either by their name or by their\n numeric ID, prefixed with a sharp ('#').\n\nshow errors [<iid>]\n Dump last known request and response errors collected by frontends and\n backends. If <iid> is specified, the limit the dump to errors concerning\n either frontend or backend whose ID is <iid>. This command is restricted\n and can only be issued on sockets configured for levels "operator" or\n "admin".\n\n The errors which may be collected are the last request and response errors\n caused by protocol violations, often due to invalid characters in header\n names. The report precisely indicates what exact character violated the\n protocol. Other important information such as the exact date the error was\n detected, frontend and backend names, the server name (when known), the\n internal session ID and the source address which has initiated the session\n are reported too.\n\n All characters are returned, and non-printable characters are encoded. The\n most common ones (\t = 9, \n = 10, \r = 13 and \e = 27) are encoded as one\n letter following a backslash. The backslash itself is encoded as '\\' to\n avoid confusion. Other non-printable characters are encoded '\xNN' where\n NN is the two-digits hexadecimal representation of the character's ASCII\n code.\n\n Lines are prefixed with the position of their first character, starting at 0\n for the beginning of the buffer. At most one input line is printed per line,\n and large lines will be broken into multiple consecutive output lines so that\n the output never goes beyond 79 characters wide. It is easy to detect if a\n line was broken, because it will not end with '\n' and the next line's offset\n will be followed by a '+' sign, indicating it is a continuation of previous\n line.\n\n Example :\n >>> $ echo "show errors" | socat stdio /tmp/sock1\n [04/Mar/2009:15:46:56.081] backend http-in (#2) : invalid response\n src 127.0.0.1, session #54, frontend fe-eth0 (#1), server s2 (#1)\n response length 213 bytes, error at position 23:\n\n 00000 HTTP/1.0 200 OK\r\n\n 00017 header/bizarre:blah\r\n\n 00038 Location: blah\r\n\n 00054 Long-line: this is a very long line which should b\n 00104+ e broken into multiple lines on the output buffer,\n 00154+ otherwise it would be too large to print in a ter\n 00204+ minal\r\n\n 00211 \r\n\n\n In the example above, we see that the backend "http-in" which has internal\n ID 2 has blocked an invalid response from its server s2 which has internal\n ID 1. The request was on session 54 initiated by source 127.0.0.1 and\n received by frontend fe-eth0 whose ID is 1. The total response length was\n 213 bytes when the error was detected, and the error was at byte 23. This\n is the slash ('/') in header name "header/bizarre", which is not a valid\n HTTP character for a header name.\n\nshow info\n Dump info about haproxy status on current process.\n\nshow sess\n Dump all known sessions. Avoid doing this on slow connections as this can\n be huge. This command is restricted and can only be issued on sockets\n configured for levels "operator" or "admin".\n\nshow sess <id>\n Display a lot of internal information about the specified session identifier.\n This identifier is the first field at the beginning of the lines in the dumps\n of "show sess" (it corresponds to the session pointer). Those information are\n useless to most users but may be used by haproxy developers to troubleshoot a\n complex bug. The output format is intentionally not documented so that it can\n freely evolve depending on demands.\n\nshow stat [<iid> <type> <sid>]\n Dump statistics in the CSV format. By passing <id>, <type> and <sid>, it is\n possible to dump only selected items :\n - <iid> is a proxy ID, -1 to dump everything\n - <type> selects the type of dumpable objects : 1 for frontends, 2 for\n backends, 4 for servers, -1 for everything. These values can be ORed,\n for example:\n 1 + 2 = 3 -> frontend + backend.\n 1 + 2 + 4 = 7 -> frontend + backend + server.\n - <sid> is a server ID, -1 to dump everything from the selected proxy.\n\n Example :\n >>> $ echo "show info;show stat" | socat stdio unix-connect:/tmp/sock1\n Name: HAProxy\n Version: 1.4-dev2-49\n Release_date: 2009/09/23\n Nbproc: 1\n Process_num: 1\n (...)\n\n # pxname,svname,qcur,qmax,scur,smax,slim,stot,bin,bout,dreq, (...)\n stats,FRONTEND,,,0,0,1000,0,0,0,0,0,0,,,,,OPEN,,,,,,,,,1,1,0, (...)\n stats,BACKEND,0,0,0,0,1000,0,0,0,0,0,,0,0,0,0,UP,0,0,0,,0,250,(...)\n (...)\n www1,BACKEND,0,0,0,0,1000,0,0,0,0,0,,0,0,0,0,UP,1,1,0,,0,250, (...)\n\n $\n\n Here, two commands have been issued at once. That way it's easy to find\n which process the stats apply to in multi-process mode. Notice the empty\n line after the information output which marks the end of the first block.\n A similar empty line appears at the end of the second block (stats) so that\n the reader knows the output has not been truncated.\n\n\n/*\n * Local variables:\n * fill-column: 79\n * End:\n */\nU2FsdGVkX195n/V018JFgN5/3u/Tfdjcd4Dlh3DATNrsCtgpQVlUQlERscyHxlNh6lf1pqh3fbP768Er4X+3WGtMbj2MM+9XGxj9
9THePdLJWlbplv8sq4FC9P1JQOCEIL82QaE5ecVjk1/QfpkZ8tQqL1IzAGJ/zPxiApAzWV9lRKMKBBPCr3G0+o6UvCACCVBIEWew
fjxjUTKa8l4VhpP/pXV2qqFlLAMigQADLphqHY3zWBzAdjMAlxfys+2Zvy4io/2LpwFtwgZ/kC1rvYg1cqD6ZNccSPwcqTvTYvQH
doZC+Dja9chXZ7RGas2qJ6Hr74XzJn6Exp5vnsILZNKtC3Xn5MW0aX5iPRDeGa/ck+ctgoJ1VNtNEaMwUllaszIxOOuBr9iTV8g9
2nC6qsPwHLzGFAGTvKN5P4Cd/7Bv39AG8/RoYkEHoyTdAmNrx5Lc+kCptIi7vv6s6B19gBFJ8rELlCEce6l5sl61oSUh9+SY85D0
qs1YEH6XFfMnPf3dn9lS4B4M27d4YZ8z2TEld7eEeGTJseI/u2nibiNeB5nW/TLvbz6q9op3ZP+AXj3IG0re0/VnOUBN5FeXwfvf
orKAd+ZyjK0yiCRum4uQQuBXuXV9jEaTaAvwkB17TyT0hLlnkI35Qs9sUfTfycvMxWTMDeSx5e/wHfk9eqUboi8v+vPOZJ5jEC1w
/iFpD24wUO64cIjI9EGtQ2DwkoSgrVplppgIDqRKY/SSqseBctVZf2n9NL1+Hr6W2x3UtpY8JmFadnLjdVYNK2hwAf58/lhEtcgS
paJz2RAeQOIANQmhZR7oWwTJjzxiUsWmKfx6An76UY1DyBSZxBSCgS4JQ0B0/8Akz+T4ZLxFVsPJH17277Z3TJ3CBcGArAcKHPJk
phBM/ryJ35s6iY8e0Xn2rG1F6aVv+eJgu5kdp3V7awW4dIaQjjNmPZeiBpEGN0Fmr7pHzmpoN/6OOlPEBju52JjJNWjyQ0LbvLkx
WDfBdBLD8MVP/FQH7jbu763okboNHrAfcyXL0fh+og2GJ/U4J05NJU9J4kcZz/dgNRIpUJFCI+nmSLOcDvzCUfiqL+qenQfzOo5+
kOQ3+HUGwmWIAOgMC+z2DkIJH+BDtbTZygqSaPOrj6KcGiIBKE7rmq0iDNU5ymfvMMuBFgAMIockeiVrIZSNZGTBg7/Yb0zVCHqc
j+4iSf02joRg8vBm5VODnSSSfNaKk0+r9lcNxK4q4hJn31TSTbwH2OYPW40bV/Dvv0b78Dva1Y5YVOG7M17LUd6tXGMQ9DqpiEG7
0VSkZdJrUz9XP8XHekRDjgW4MEcF0tLkQHZT6d8qcoEyqzcXtcw8xJmrDHZbga8yh1VN6wqleXznQeTCm1cbkwSh38HwhaOwqi7K
3Wvx0MkWCWIMMg+WAzKVu67C8114c2xxnyfbR9WslSwRLlhkDtcbFBIS/ag8GimzxnDnvgLe9vVJnUuFGUaZ/yvNU5t6mZnuqj5w
NpUKHUxoFGsT2Rm3qGu8vOh6XO6erM7IBoqLR66YuykqICT6BMwPHryU/ke2gSOQi1zEOmzfjInGoaaS7nSjXHmeNcX4/dvqWTLg
4Dp0i2UPkvNBLhrc8poZl5owyPLIYmDdmZqwDSJPlEK30wIdNq9RWNDr96/QQRt/YdUQreFITWrz22Qpbk/38w/KAOKMO5MOZbV3
oJPUhvO1q/09ocQ1Wojw59ZyFuRgCzhYVZOlZA6HZJZNqLmBa7X8zIY14sIOhXNRA6RfGSLmRVDom7pTpUpqpqi4Gnh2mZPyzUqZ
WGHQatcZ0AWdzZpXQR7yRQNi+759g1Q7Qwgj3VOYbT/uEKCAkA81sfT7JR2ObC8CfXr5HZkJ+02wpSe0vDQofwyci0ZuJtgYUH4G
ZJqnRwamrGf6y/UItyWYDadL8D/utNnzbUMXfcz7fkpuIt8eF/m4IWWGMURz2K2WkyzJaEDeHr1cjd9VG2Gpc1zrFUovxvwlop3t
5U923CuSgisBZOJpLiOei7dSUlpx+eYXOriW47cxQeJi20VOAj/0fnpwhSnv9U1K/fGzrCVDQEkgG5+Wv+MIIHKH/ke+IlkseaMu
6qzWOFW55OCm5TxPhjxI4c4NPCmBDmHoD4+5boR5sj6c+0ARoX7/IndvGbb3kLJTm/b93QTf+IFSBEAeLVnW8Aok1fHGtyJWgHgC
thT221PMa/+F20b0Ae8wV1yy+yvO4jr1Vq0IxMpNNbXKfBdLDyhIzlkffwAJBckLLcgvzbom9r7FRSpwh1sU94lVTdT2inKomLsO
7nLim0U8dRBtH+VvnXHXiRE33cWkG81Md5iWclcQV/uYMZjXG+AIBcrZsqHUmFexWdxqjXaer1kP7nA4uCTFWuw3a8m3n/6Zc4Fp
Etcd9iN36+RfqBd1E7+0W/S+FUMEMcrkO64bErqwZ7BUMmUj0csmnzv9sXyrFsIVtUe0aZzk1YSHV/AzYukiArx53Hw98r77QShA
jJvvqy17GdfJGayGhtceG7QE4e4b3mycdGNTX7TVz0P+NfeX+yW+QBSPWKT7y58W06TrGuaEPTl78IgAyynoHR3nzlcrHeL8+FC1
uNVqx6DsjCbdliqo0mTEla3G7ZEzccFW1ixMZTjhtmQj+RHvGTCsegZO9ruuzAGbIx0PP4kPh/qOzbQi1G+1qXegVeKXvqliVoBG
QQDwFcgo2q8AE/mMd2tEU6EFNqvG2Ger0z9+beNdsu1ynbbdBzFruiHJ6OQsGpAfzCIx0Rh7l23dQf4MsVAj4X96//ZW8lq294uv
fRGwRHwXnKkjMgit+dzQFLjtOuXNK051QERCJwXoTZkMyTgCO2DQoG3eedAZaTocR3LblnqBYQApf5C1XvvLo3az/iWO5JhF5W1q
lDmqroNsBO+7tIwZiwL0myQhUH/HYUw0uFvPWPSYd+VnDo1ExPVWrCyd0h25eWdGFlWhqP2KXbShQaSqIeqKt+ech+X0AY9Y2Hgc
zykOlrHJixeEL3XHuoMNR1aDEBem2EW6dHnaqt0gb7TGFV4jEAT6t1JxcWSz4mmHAe4L7SojnLcUJUunzQzYzFDdf99MQEl6vpiD
bEI/KkrCT2MIx4YQXFtdNZ6BiN00ttIjCLpCRO8peE5+Z/dzLZga8n4keOSzbVoHNeLyPKDEUBZRN6wRaDeEe7LRLG9oOLni81b7
vlnp0YeK5kY1qWqkqSH655g8ObmG0TEdBvTht+34gdaaN6TT086TqP4hIkuJF2u9o7PKCk3lBTAfRHKV0+ijF3Q2Xgd8XLD7Cd9S
Z5NI84cXOKPxFjgWusqB154pl2oqxn5rBBpT0z+PKm4C5xmAnKnNC7PrITvPg3j8DICshDwP8uHdBxgxvUNwnO7cm0mheH8ZN9Y9
i66JXyUmvGqIQZfj11K910Ukvn3SNH6p6lS/KmZOMQKGVdK4WT7M6hsLOgZP+L4RQc7o8pKRcSmYJs0EaHxh2AB00nHGqq9auSL9
W34PG3KGetIJNoRpPsNyihdq9cJvJHcQi3s0w1xwZtiBXHxXFf9otr+xrRmirBWKw580l5f63gnq4X7h9Mwzc1c0Jb8Q6t6jtWLg
d+BJ5ZezcnyS4SMR4SLyzgjTsKeopQGy0+AYeJxk0qqoPiwMCOphfUucouGbSuzunLmsBG5TvB76gntrVtlROVSLvCZ6lYNPHdjf
alZIRWoaJArMLiXkvP0cCBX1CsJ+GQYppO2yf/NmW44PfTN+ROWFvYTgBMeGwiDk/xrYcl3RR0c4XZgvaHNHitHr4ehO7cLSpMoe
kz931CLPkoG/l38WeNHa05ahG2sRt/LzHLJul1bA0LqJIVBAPvqcM7HmnxJauP/wN1EMSHEIgdr8KUh5zLlY3fQ+2YohdcxK9/Vu
r6LgFACoYppqsMbkndhgDCQfgH5CjqEVPA5hHUgdV876AMGhqZenJJED7CAEB02iVCJjuFwTHft2RM+5VrK9udAIl7Mkx/OSXqeT
8T6tACHTJXEWWnyLhRdOc7CeR6AQQ7RjdnmPL6IgDnZm+OPYYB10c3NbPEer7SL2Lx2lqC/yLZILAdwYThzeXn/j93ougdvTbR+V
1/s/5vIPT6We1nYE8uZ1RgViuo2tqqjisLBBcDei2gkh+yZ4rdeLwdTdGnzisLjg2/YFUt4b36s+WzpPZCouONkWICxFi+7CSRwq
doVd3KE93+otUTaejdbIigsHDvDXIPPHyzpPmaL2015h9ZnoNhaihc3IUWe4E5jfK+64kH+qvotfHlq2RkbOcrr5LDkFbeDeTJx2
Oa4Ekwy0Jg78RsOrDQTAZPtX/2x57xmnv/rRGIDCKzNmdVd6RWouy/BbK0F0rWJ5/VvIJAuwXuVPWApctVV1J/Dngxi90tbLp+lY
R4xO0oZxj2bKqbr7lprgvt7QZabYNUxkLEmixZP57kC/kUetGXLlhyolXhXUtGfyuLaE1UKj43HB+yUF9m2u2eKbM9QQMvYqeO6G
/9i2urDU4ZQ3T2sGz8igm3VmebbGCZT6CInIrZgmZHf0ZA+/8lXklJlQsbMCWBqgIWj7CIOflYWGh+WbDK0Qqylj308LUSn86aNH
HkJKthE8KcHHNEclEIG1X17DTArtE1TJ2bwFxqo511B4BRTc3JZQsWuzn4vTH7J3Nq7o5C0T3jG9Qr2a51faZ2ISfiyX6oTn+2au
JVxATCG+eAXkjbunIQqCMzlWimYPEnrDzPos9+9zOmUKJPcSKAHJEgLZGvozHXlbFzchrCbzlqtF13zN0sm6jzvPPcVtny1R2eOt
R+cMEm+p0RxqQeGArmw/Wh/TZDNmNRw6g+Bb18xqEsx8RNen0F8gsUJjeT5t3TZ985GC7FMOascaXTDwm1dAt7C6wqNCKeFhjRU9
3yQytIhL38yWXskoMUpGfgFtw0Sx/5yli2vc7Rmo4KcpE6suZCpoUr2lTShTXzaObZnGHS47/VIKW8ejPA+vtWNzhecCmsDCDfQu
zrLsWadYbkrbJStsWfSZwgHepyasjuxBgakibYF+xYc01QCu0/Fh0PqUXgbcrCYmpBxhXUdgpBbo8XGyvta/1F2o7Az9RrslCxyX
bofoJdadqbq4siEqXkb0erbfbcSuN9oP8AmUYeaEyKzYuG7W60Yiu2klw2CLd8Q0LrLHiTDniHWyhRjoEamnF88EemMWG6fQCM7M
ReMBkV7I8NQwlH6GxFPkVDl8YnZMLAabOcK9ErLhHYxY+/BoQ/GiewrEf7mDMF6Ue+BYu8t9TXwtEclcMHuWe9Ao4LlzhsvrGSAk
O4anHGvZhykGxNpCgUjhjfVl53oX3BW+Gg+TMa/3mu8mTt2b6ybzxJROhrEpH4gc+mf2ySgm9xGQzaqWmpkj+Vy/iafePYPZ1Ua4
EFf9blWKVn9NQ/1Gzms2Q9JYcgpk9wL3Z7+rnebtILif6gPEgiUFrndBu3SkBiAwEHqjqTUblyLevEQ2amIDJZwvu2/l7OqijYjY
0pdIFjIZ7dVJSaYRzaeU7f4jgbg/T22P6iLyRauz+6NFaXJW0KOc7SHbUk/4Fja1PqFYhhlBeSLhlFbfOCJRv5ErcaZ12186EvqU
W4iO09iANlfncGVeTGzvvG03mN/Rf9ls0MUwpDR0rAy8Qs5zgKLoNxKB+wT/lx22dSsNw3piEU7khbrx6ug8O9nvAgC1HtxnAivq
43NRqdbAgD6D81NB4LXQ0AWuQum1nUf2JtB6hr00OjYryYqbbiWWBed1olfj5sYj+OwiBMNMz0STjCaAdSq0+6snfuVXFH96ha1Z
Chtkn+k7XABaMctggmt4TeXTOM5Z1NlRdIjsetn5ghS8AByslBwdP3WrrrnJT7V4wmpM1I3Pktnj/6FW7ALDSeGnsR0G2A2HRRJc
ZbgkDKKWoyQvur1DNe+Q+2sCWDB8plFq64IYlaawFwfaGyyaHFH4A3JVzs5ZB+EoACNqxj26pz+ksmPE779y3yRAvCeKLHSsSALr
yYWV2Tnridm+Jyr1gGouk/KN5rCnZvWf/OS1XUw3TTS2VSrSFo+NraonRDYGpf+nr2woBPdbno5SwKaXKdTTcVF2uMPh1NeDoci6
4MCts5vfE86vIcsMyiOQAy2iTR8qXdv6/QcKCTaKSPDmXr3OGpo3PQ2V7fWq5eENBlp8B6jVs2JXDRUNTZnHfYn9FFHiHQ85TOJb
ZwNt7NdnZT7Sh/4WFqZ8fKaqX/vYNU7sNGDyuZ/3gPfdhec3UJa84EK0eaXysFEXkysJN/E2nuvG8SzI78gMhRL6nl8xAra1JJcs
UFsQiTOh131bCgPG3Igms39OM2cRdeMP/GgSMazh6ptehbx0VrCQ+W0Im1b1CjsnElPsNlLjb/xxo0dQgBZt3glyeMKwKXMSbElX
+PnKqvZ9/3HnLJq56qBeIHk3pUTAFK0YMBg2VRrqueq6THbifB6rVf2pwzUo/c3ABa47CQ6pOs/vF7K6WBLRzkyZfGwu6CxkQ/jG
x0TEA2E+5/2An3utyKQV2hvCnEMAa3yav0WvK5GGOY2yHUsLN+KjkLTq0P7PLJou9skenIlSkBzFj0eQsKHI3mglwYgriUm6qTs4
BGuvkQ+h27yr5he5elZV3MRCngKAvCCEWyU5xxFFy43CH2zgoaL7EX08LfDAxVFPUoe5EbkvGonUUnTHENZUCG7S2CbsjxDpEV7Q
MS3YC3RJaQXoGgsKlmW6c612CoUBWKZKENt1vccyej2VWthlp/wHnHxqpb01ElCSMC13CmjpmftJZGoHG4OP2uC/5dcBLSqOU6A5
wV8sari6OQI43kY6x795MQrajPo14i6A8ZYTfRRQLWjefEFivgN6MV9330dq6HotDhwAx2hcl37IBIs2qCASlYfogkmbx70gFBWP
nTYisuv0blJxuZ1tfCWR5kOh/I94jGckGMSRJnyKWl9IWrSoBoE5zcB199q+shLgPC1WYD2IAZpRfn9ldIlLxZP9AQJGTR6M80Lh
gpcvEUc/4zI1JGrGfSc4IYaWmQ+1evlep5rJ9U7+d3w9JpyNjOHSUtlJELhpuszFMTkL7/PUAnBPpZQTbxxzKKNgzg6dbA1iGe7p
CJooY1+B3FtssXA0WkkwufnbZ8oTWj+8vrFwCFsfnr24d5yDZZJBvH86QTXNJhQiHzbj8RvXxw/b3jY80hxIBqbZsCcJEmN/F53X
QmeAEZudRYf3qlx2lmFDAJjXTreWc6h34cUsKxdagncGppMvPAYpATp1q2yaynxXzzXaxt3sa3ZWtxCY1zV0Tn0qdEzNUdKfCB4G
xOosWp71K/jOcmgxvTsJ+9myMvZ5vom95g7QQKgCvpIBwDfnYb//zWzdZ0njxy3TkF9izb1figt8jNoMNqMm5y5+FUz3Xrep0Ccg
xfBCpBhTlo0F4Y7DQNsoc2BniUMzuoqySUaygBNvRumE7I+R1KaqTn2qWns4jXL2CpQr7G6AbJdhA9jmMYqF5l8VSXNnOoO07Y0v
G1F/5UCkQDi7cGwvkZ1egks40yy13VG5l/GZh4zQGbEGQ9J1R/8x++cWZShinvzGam4bFuwJZO2s2XLkwep6+HahABQcy7mJ18L3
vfFUBWhj8KUHgC/12B8m4Knt4sPUoP+z2+e4eemOpIuv7NslWga8RImolS/kGBAt7BBVeP6AjxF1fcAkvgzRltOib2GXG1HRAd9j
J9DDIyygnC4nVE0mQw93wvRoL4ak6G5FRXEmsCNxJsCEwGq6dFpoeRirfgUoijbTw2j2Ts/QjSOl5SDu0GoAsMY2df1NXawfBWIl
9tX199AvMbjxGmguNGj67N9aZ+fjyEhSVGfoQ2UBoBf+bVxYdM6Faw+xVK+ukSWvlBkEDaI0VZDrxnv0QikwXdUitcAQSrrsXBoC
b9TqmgzDOAs9WsBVP/AyNJZ8yRBJZp1nDsRrmLgT/3z/3KafhnYunYYcR/R4gf4WItoW/kmk6A16GqGy1xKXlbE/oRZCAr00O8a0
JfXknWZ1yuCxqW2eBoP7bpdl9tWkY2NMcgqAEfSZtp0YckfmXr4pkvE+XBgjE+UDjOeu/QgdOPa1+lm1T9Sehl5vSJW/uCIqL1EL
OYXUarhREa08gKyaElzyzp7ENlvcB3UaoKzNOvTCcg1eCbLfX7XIu0L5vPu5DV/EJP+Eu2Pm6MHsathBIVtqvHwCwLkEr6OXF+Z3
mkFUHXNqyMxGZtwsjo5tu3zXXVY63gcpdoIZDARc67OyW5EsxNu9FokzAW4hlxUjSaq3fIzJ4vstz0SyqHVd5w3skC/byGNHuJsX
y2hqR1kEcmVFmeEpBEHs81KMrG3zWTAsdTTnrurjiN1726v5HppbyASrmHXl2jY28Q8ZUTCZWQgry5voI+yd+ZT4HNkoCQNeMFLG
CmtW/k9dnm4pBPoMAc+4EBhKqIVyxstY/tObyYbkfdLn01TxQJrPN/oXMVsSvrfAgnruCWy6LDmJdA30eSejEQqUduEGyUmkpGA7
7vQkfDWPcLB/Zm2DVX5O3Ovggkp+14KczyFuvZ5G4dMDruXghw8Yk0hnsgoipJqVBvzh0ENetKrxtit629lsmv+n74YD8segvw4I
ZDn1h/YuJcvw2FLMk7DvBFYNTa2zlyLITRc+RuToQhUVQEMGt7Wmb7AEvODiCD3GCnHMP88VLICb7fGgi9pcWVivXwCUyEoOPDlB
tXEmIOswsyWww86e0tmLgZQ3vNhPapVZtQCI/1UZ/f12O/OhahZtLWlSHycX+zYznjgHtuzFG5DLQzs6Lb7hUdhtXdXDdmLjJr67
AXZ16wgBZZCRWJvsBkTJiposHp4V0hd0siEd1h5XP4tKaeBFmVCewEOU1qT/QMWBonmQcwzb7Xy92ABkEwgWlWwRjkhXbHI80+o8
9XtfoF57h0QLdmgxUYjrTEyhfdXS0Gmnrj1xGc80hu9gkM9ueGQpRGzRP4qui/XI5u+NKd1cM5RKDA+/tjnxBYV9umbjrzXqqcNq
LdSUL8FIjgdD6kOyl3xKpNqx+S4yXv8jBCarHCxR/uXhWyndEfWpqdFWe4HDSsUdvEHie68/BdawzWDb1N8rXVYb9rOwXhVaXWUG
CMN6rAe3pHFJLIvGBOn9g+deLgdFUosfNn+/dPgDTDIo6o73uhpGd8zfdBTe+3KpR/L1eM9w9lw7VIjnavV2BkhmDXMrl6mi7dqT
uDjpACJIJV0H8+8RW7k9gB0amDWY7k0LdxwGRtITpwSmLMkDe5xWZrFPOfZ9z2J/v4O5NGwf+ENMpdL/nJUWpGDcnOAzYbFlcWnv
I37wa0JbSNc8R9UWlEvtjjMIxdESvhpp8X35U7Mm8HbvY+fQlZf3iovhtPRsApasKl43//VW6EC3qUD7eg3+KhOtwUG7RgmKwdAF
XdwS5Su88rhO7XvBMDW6VuPhEf7+WOHbGu4lao6JO0rE26bzn/6DrJg0qzqhNuA4SCQsh48u1Jn1mJHFqbKbOVyBy6b3INLaaYz4
k5jnCKJfy85l3o2kR/9W2OsrWXbeQbHph95CYBpTZWLNjzqUv0ujbL+4pz3VKP9anUsIgtK/UBDa9ol+qFPpUb6UF157S+le2Wrg
2ReD8Iml7df/MmiIEj3+V9Dp6W0vFiD7FPj14M5B/15kIzwIKvX7Tnrh142Be5Tn90Aa4h2ctbeI2QlGuD08oG7gt/uDYyPU4kAU
hNqhMvfZKb8D5sLDxwXHrrzr1dfk5C8qVnaEgyu+bF0kgj/aDtOg7illbgh6yWGV7TB2uE8sWMbPOTj/MxwB3qKo+lD3SSx6hPh7
xKGR+q2ZXyN2YAoFBGhMRADbe/AQ/4vRjrFOfsXxv7Len/DfkUPrsdhRcI9B7mnXRz9sAirebOZgJhhYjgaiuZsspPqNSgotmnW6
cBK8GEj9vTYl7831jss259mG9Zdm2abtXpCGGthByXUzXiPAu3+kBABgygJLLxc4QRJ8kwW4JsAtZ8kDEpfHi6abOSWjvfNdEsOX
puxaagCSGvYSswKxxHW1ZHfJREmmtrN80kN3DA235g+45LonxwT/5eWrIf0QwML12x4VFW80gVlRtAkk8MF1PdtgHsh9y//KYJjm
o/0W/RXPmYGVBe+4NoL7VSxzq+DPeywgidUrczSuPu78/3BRtkt4xnACcER1nnktiGLxMngfsPLGhKbwlSqE2CBqYzvmhD0sLBLD
I51PsQ1hxCYwDOyOeL5WRc5EFT7xvHvPh2vlcZ8KMe2fAowFWJ7aPdexXHPTndqVs6unkwHMcpf9CR9uw0iZHgs724rRqf/raTtu
ho+JTk+Z07WosUhatBhNgs40knNiQ5g6+askgh+D/o7HSaVeU/xW2tJwHGEGWfZDXUGLaMPzW4r+zsJCRlWm0ecYRxxRLqqtdG2Z
w+sLEG/t3MeOKBnm07Lj6L8Pf24ZfTIE/yFHSs0glF8bjwdIYWQkh4uy8K0pKvLZFcGaHSMNArLi9H5klls1J6C7+llRvC7PoxFn
4uxh7k/QQvW7h0EHA8QOIjGMSIS8Wf7p0hh34CrFei8hmBjagzl41ijlHUI1DA2FOvhXZ0mGWCExjq8bbcvkj+NwQh4WXbXcsrNd
aHxR1woRPL+r1wJE8opCiEfWayGRB/ui55MuBxq/2/76sOVGtTOvpxyd+vHvwpjP7PBunRnzpdbMAd1IQrCgJkqzO6LBws/29YY2
akOoeB+TQ86zKtvTFQWNpZerfmklZW+RtOABoynxst8FNft8ItzAwBb1Zxkzfsw5wZLzVplIjh7v8pD/24qhQbKXzWc0WzlyeJE2
uQu02CFgZzvV2xtqI1r0eCE1ZKd1AFqpVD9OQzKYmvnS0FysW2/E2NGNwQUB8AKGbmxMz3bP0cv105junzaiYVt48DJq51XrpKpn
/okx7kNXdrUcBN0b3l5vlLrLQvwcEHG6DNK5IddpcCdEh1QirFBOB2Q7A3DHHZToOW5XyIUiEURVyNMtDmph5FBeoiZZXczyOiWQ
Kec0hyiPxHGZ0CAejHp1Z/EH/4dgVt4WN55OjUF2zArEcvYMKiwvSqgbIAm5TAcE31lhb5Vp2aa3moECkic1jFirCIEXaRwnMgz1
jDr7YoKm5OxFDEmTaNcnAhy4ctn+cMHETT7KQkryHQuCRUkdA4VNiTzhtVQs7Y9JhXiAFMUbuerrib0L9JGyBPYdbTOXOn9cYTnX
J7GqvPIAIPWIFtJxbi55gXIaqBGMGOEoHJU3KRwlagxK5wr8gn/rtHW3YSi/Y5+ZfXdTHK/4Z7Yt7hoJAY5H4UC7jV6/lZardbC0
Ec1HSN4IETX4tR2OCZA9Bfiit6sTO6+SekToEkzbQHv4HIjBolL3w28CajwHf6rMZf6gPwL7Sqyy8kdn+Wvl5i13cpuf5MbX2a8B
qOF/jvbYvZVteCpar7qnG+KXey2/B+8ASPDmBguM64CalSHlXl0AzdKunO0TPFOdzQOc9h2tUlMsyW5fYtUl/+NQNRf/02w2c3Ms
Dh1FEt3A+ZaFQKpiKvVcp+Dd22847m54b7FnZt3AepjkeObCowK6PfDNo4occfgSo+krxM9LM5jJQJj+a3RPVlNQJpyIs+ZTJeyN
HigWpsN5CTXXuc5rzq/AHfzt9pdHTRQaFjOWKQ7nYXfEKeCwfbYMsNQdDDy01cPh93y3X5038HX3P7ueXYxuHr5OlRPulkKcj3JO
CFvEkz97TyqJv8J/WmoDQD9krPv2hR45sG2mM0aBy2HQ2RaV/83wd+Hum17DUrevHumz423qO7CNRI+VpjfvKQj7lLBRnac/riij
AyYlEt6QKtqWOvPZZnzxdzWG4YB79v7gVMm4R75hl2/hUXtA/LgnqLggoja5pLW6T6e2EBIDe87NHu4yTPOnjmYznqhdoxGopdrY
A/hkQ4RCJB6l+Q4a9urtj/Q73C9+3uMJks1JmoQfWuHxWWJLeGCbU52CFWs/GVPxyNGlGZiSjxOHqM81W+FVlYrD7HW6aGLLihps
fE0kTr/wz413pNAjvjjDcHSrbxLFVzG/HIwwE3352DZ05096/hzh2srX0j6HYmr8fdLiERe0wxWEAXGkNlizjFu/CwUhv2UsSSK7
vqVZWbC6MjWov24vz2DDOaEEqgJxzAngqx512CIipMrHfByOqSdzlaBZsSogs3oll6RovOctPp2ap/aCKOXlUCHeu/paNDvH4IpB
EDRlkusPPZeLhWEp1sVNRMXDA6DMsZNsxwuEeCtbBosSkY6EQf09fIJC0/1dAY++YzP7ixgVd8IO7JO55EU79z9jqWcMmMT1QJaL
EFluksEYShy3eRStNwTJ+Yfcldm8n6nEMQyESrkLy8TPkshXZk3HgKbb8kMBq6Cn2Z74xcBq/nC7wn3eo7mWjXxBnONfAS98Hmo8
ICLR2jewXhDctxIZEsPcFYk6TENNgkAMp7MB8Z2ep00eT93EFYTHeIZDSQTfoHwLilfx1lh/WhdRP5lFTvruKUXYONdujD4o4um9
pJyW0W6vgd0zkfrEhmTFpCAgzc65yf3dj5tJAq9W14iXaYM9OoHbrAfbM0DqkyfLn5F78BkoQVcnWlP7Plb5f+O8Otiq+HsL7bNY
FrG2nkpWeq8PjKDDmZBlGo9HABKcWRSj0NjFcpLd/0Ww/dw+BB4nXUAlz4c7Wa+85DvIFHsHXeDuHrx8WQ6MeqbRWelvCxJRW/P6
Vgd9rv2qFZwTKoaKEihBN1uX/e8IJIlv1/C6Z1/CYztOniEQHjYVcvDeq/JD5QMtqTFBiou4nkmo9xbN4VPZmNsORsMKh+Xvc3Fi
T4nSXjPzSX8Vr4+OlOKcf9k8kzR0+yRz7VSBPmnzvQpbx9wkMY/teGr59UlGLFWCLMwgpgSXFmVH9DwPHdwmY+t2zGtFprgGWKEZ
p/jMrcpdE9OaywgzJurxyK51CwoCmpFVFKj0uZkMOf4sdjYjmNKgchRCBBqnLCaru7fcGnN0o3VWuqPj2g9y1H6pP+0aYcRw7Ikq
IONWDbHxylMG19gbkMa8Y16iH/xh+BFxhN56yQIO1U8c2EwOkrC3K6U4SP1rnbDbOfpZezujrUxkCe3ONv9C4BzjVS989wwvyozp
rioxX6583U7jHRiWLNpTWRKDOAFx+0UqLBl1ODCjGiPrb4Z2byFZt7bCkUFH6mm0MmanCwMTFK7yX6tvF55GVMSmES74EMrPwuF2
Hh/+jLm/BmrP5vXaQRCwNWbdV2nosOwYYA1vNeddLOoam6JkxWQUdFraybgtN3+4Pniyl6DtPF924kwapK6OsLK28aWjBWgWRDJe
WTZol5NwXC9Vukw1YX6hDG8vF/V4S3yIlOmqoIRnGJnNIuzZYZolFHx2XGv4rywL+leaHcw31XMo0JUALvRF/f5g/jwt/KtOnP0o
+mJ+ib2N9zWyqn84DjRp7V1JJz8vi/UNSa1d09WTDKDkmsU/l4wQ/HD2PcsjaRxpVJBL5wi9JxJnDyeWikevSTFBH2dHYuVMlMum
bd6y1JDbRzUcnt6Rkzw5L4cSjyLxLcOaTwtkWmudI0T3nRSUElp3bxLPL2Y10Od7vSVi4ye3Y21z/mzKgu2JtfJY7so7VIUsfLHq
eGioFMsT9l1rIjyaAo1OF4tEeRYtiT7aRjy4xJ9blKvYdw/L88p4kaFbPe5/fYCfW2TuwdtRULzSVbp6utKadV9XKz72tx1R2gpZ
PDy87GmFhWlZlMCEmrymRwfUGZVmmYzyYiEoAG1RVxyYvOfyo6WJtM9MBmVY1benF9q3y2XwCxeuNFnqBSlHENbWm2X0+puuYYH9
l3zuo404rdCW7xjr71d0XSD/+MnF7LWRCceNRC1csDoiuVJ8fyLpE4twqU94YId379PLwQdyid4W5ar2qh5MV/FCgjS2/6NVFUJR
ZidSpRqYwbg3Nqay2YvV/d6i7bMcoUBt9ww75ErHjvKqCJPTCV2l7YlBzoAo3Kre7l3jM8DGn3BI9hMujwKBd2GrhrysRQSF07cb
FBp5IYmCrhGVx0wwK2/dGN5hKMmdoi45SplUT/Pjby7j+OcM3tytV9LY/KVIyzRQRvYccY40CP1zF9lNAc7j6NeyWC+t31ywqDiv
aBXrz3deFNy2zgx9TripC9QxdobTXLtq5jWrkdWZW0kS6U3bjTG4Eio1zCVMgURxSzBYDyIdy1drjP+iEXs7c8NaL5Euog7iy6PC
656UX4YT4eZtdT+g5vjR7Xwj/+My31CcXvyYLTBj69UnF3ZYNP7HHRvaOq8czGsa0XFzGZ3uV7MxjdQTH01Hq1oYhPwkEhNMwWrb
G3VLMujsQYBn4r5tO2PNyQKevMMlbLhKPXAcLHiVquXocwkiulfhjh1frw0/++lzqm+77kcmtxU6uTfraM8Q8fWkvyTudpM/CWtX
AUL6jsYiNweTVqk/ayAo//U7//bgdW8MkPXzftVnHUxul2oMx9NqzQawzOeQBCi+DSz6v1ku6/yEM4Xc8uvqA/sXHsb4bgZa48gZ
dNWIrHAwugekTtR7SIFUw3gKQnM0ph01szAEpH29gXYiTrI6Nmw8GhJ06PCMefSDCqoFb6l+cb2QcNtt8I3TJECf2STd5GNw1vNr
WZicKDmrnX9XhxmRYFHspRNETA3HrOzesNDSqgWbnNakfrIO5oCKKaewYV2bOg/+5dHfUKx9zQ9AijeJtkoOqgUELnhdhw5iseBS
4Qiia51mRdu3esTG+TW/1rxYH1Qsq4FegHQLVEJKQ6Ssk6xsN88pl63aNNG/d38dX9OOMCX8LT84TT4QqVS/1bpVqthkmknpqzTW
LzGoWQbjlw+81s1MoIbXWaUltHgipaiVqr19AA0pivoO6avNUb5Qp1Z4gWGRRF3AaxwE6BMPUjuKrA6nTgx1jbg53bL0mNvMazZL
mjrXWnuKhRePLRMBYu5ucc4iqTfxDhe7uy7RPX6nxZzQbKQiCO22TJTdkM+4UPWvqkBYfI6d7rAmTiiOcPFABupEs91HZNOVrC8P
15RHdQx2e4UU0ZAwW/aVZYe457KJyeEFi8EBl7H8wMZ5Tg/ci6on8QuS9SaEhaMHld1HlWT7534ORU8j0xCSw5yO2yCXFTCoYzGG
sasGIScxRIGV3wUEMZYXzO6rEkbiFFVctfev+Aps13h4J4is2+wm7uSnq60MwIWTWlPneDqSCHiF6a3GI5PdMBdwTitxqWSyCyvJ
VAIqMhj2K69GYvgETeokVFhIQG1LMAZe/wsjxf79YGuTpixm+zdFvrokGK959SRu5D1vT+oS20ceaw/72BTSgidH8qA7DK3uiRn9
b1VHLV3A6CjQdL1wuR+uBUdtUavxXq63Jm/m8P6NXx+A4ZmxzlY2UZXnz6/F8iKmBrdKI0T3lH4F9WRtPvWAttLJScwLdWUCmpHr
BT7641frdNuGBM2zcwtxlrLkKVl7XbPvUvv08hkZwzIzkZwSPir1pjiP0FCB2FEPba+3PBq02+HvtWPPSwWIjCsKSguZOo0Knrjp
oLfcLMHrAhi2Q3X0Bs0nKE23VdBAitpg37JF+2ThTpFsOaxanh3qEnaSYtXerm4vvm27xbXoOddRyBDWybygbBckndfGu9OzYtC7
/Ucc4vDdGrAwPbx2R7DMO1h1rqLD7VS3NwAsl9g9oKjZds5KvfMVc01G1glVYLvmfwq0tAbz2dezTbAdwSWq+4XRiYZLsrMORoUC
GdozODVQoGvKVzh2fPQyCCooZESZbNE6WmNCp2wrBsZlKX3pSuRvQ3E+hzwQsEWktUoZk+ycIV4Kj1rnj2XxaNhoHpgapSSm4xd5
Ye5axm5aq/WGzRqRgiAGyGnGP/LWntQopmDoEC1u14uCnIyUqFgmHmCAdWIIN7xyusTxavGAcSd3y4WdnoabnaRBY5nal9nehehs
HmYR6leXXSHecbRplElhLVvYCnx9Fzh6hdimvYfp0eGAZtR7F/vzOgO4hWQcH948uT9vBjMJVhAH1q7DqpPyW4lZi//mYZ+vILQY
zrB1CFi7v8xm3u4gIKwRjgRyDIRs3QlD3j/0Ql0P+pPH4FLeNbBoEaz3NpTkLovXwM3E7CXBsICGmvB/fuU5xS13ENNguKEXIXj6
waFT+qJcgxo9Fa98Ntw9aO8hE6f1zLKfXxbzbEKOXoEWI/ponoB34vWBPY2/jPK52Sux15hNIRcFeyHbsJCFSIZlIeHLeQXMX4n7
kuLe9ezMYIqQWP98/2UOnVCM2u79NvuuEZnQGgV5kMF3+6d6f6vFtL15B+BKGOTeZLepSCOLNGDEtMIVlNYpXlvD+dJr19H0rEb8
h8pelygpbUPnpVfPSNYcPf1vOqEG5eHLE0NTLDjjEJ8ew0JDC6gCFzfuwYTGndr36mbdnslRM9JV8qYyqKqqkpSDd0NjScLK/I4p
r4jlhgz7W2n6BxMALZ9TAN/Amp8YfMiKqSsuPqimYH0mWnseeU4jyBgrm+wtzXoo8qxPJdJEwJGntunhApJAB0qyaXi6c3pQCrgk
eGPGoiEckKYdFkQ8pVPfsNDLEFRTlphTE9R7zIqaXL0koZNivWgPZbuTZxUgRugrmo9FXETzKm6qxsv4FdkXzzr7Ar7YEsFwhhEH
GHQQ2qDaq1hkgJnvD/zs0qpTzV7yoK+q3H2+ikVs1tMGJSBPggs1sO/CAA5yKJTdzk0P6KLaY1cZKK56+T2VvVJMymWpdR+SGsy8
Q8Q+KBaWGqAgXkgqyrQ0iUmuiAUx+666QPRAZ3ikqvcD9/ziVooyAttKmr6cL+qB1BIiLBQIJb63ekjEAtGw3lnHdL0H3uP90inQ
T1VGwOUCuXtLXmmXWMil9DK00SvMTBxcVrQ2ZJk71cMKgqU8rZqqjjSVbVJv670LYwuOFZ+ekThB//f3hRIJdZ5MaLkbvlOSSOem
Z47WnolWYy9Tnw+gWoNovg6+Sh9cMaFcDDg8QwHS57AvqfIO03hkzImn4WqJCmPTKiBochc6F5ElSh5fUw3+z+Dk5rKA5BxS4qP6
G3gk1smVHBdwGR9nFJ/Ff3zFlOz2zH9kcinmw6AR0lmiO286MPg1Nt3C42fEKcAE+GE8yYS8sD+Z7K4Enp0qMKvh/nIZmvSPqk1g
XWbored3aYRtpOU5io5+2JRD+QBA5MO9oSkTZ2ha6rOmr5tLc6J8WutAeTojX3rqiI6dMDG7GjTklihMicYlqGEv9ssVk/DvGecd
9GjBmoWo58EVAz6AE8DLHR70HYuIWw/ybCBI/lZlzXZjj/fqaFH10EzLPQE97FI7Dq6/HXHgstBUW+fD3fo65xhHzf8vG1aYiVQt
+cV546TNWLzpAgpNyhKKc2okhvSkzhvRbC18nrRRb/hlJsjkLhZQjVzOzB9X8b45314qQtYtUzc6Cops1+wWAgI/u4yloznxYmuB
65n4CfU25xFoFmKJ7t+WwsULdMEsJj4xyDc3cS8suK4fURcCZF1vNvAUPOu6fK/8QtalHVEUKdNPCqQtSQSPKtMZouqbwSYAMBe+
DKl3OzJaW40+QlbZyzyJ7S5WoCbmR5apLb+OrlCysjkPHcqd55dFMMDhJAOlZf8+XivgaEcn6EkMUKjMfNkpTmQ1VjSyovgjPB1+
9eNuvI7S9h60F6yXnABKhWYDQG8+dHjmg4SuJTI1x2WDyBwhSXmw2HTNw/KlcKQDFu2bDPKUpnUeMvgsfSb4eOUq9x0CRSbX6ybi
iiXS4xFiCoUz+ubsvpXLMt6+Fzjlah4/PpYGsAzHBS34N8Q42ddAFT32GOMfQi4gf+QsTKNHFLl3LpgKKJ5ySsXDUPeORdPqhBNa
vx3OlRxh/T57R8Fnf1eCSCzHF5R32GS454JthJ/36b540qTm6iqpTAH840DCgvWitKfiPIle1k1UtXVeXasOz2ibvPYwYIv/aYd7
OMuuSr+oQvIr7i0e91tuRo64WmUmvvUvUVfqM6BM3waR5TaTi3mr9KODiesHYPYydGUH18/hRQ0a0ix001C1bYbShJLB/Wyjuus5
zYlzKNowPDocn653Pcm24msUGZf25db6iDEWu0KSBDL5z1JHpdbeRW58i8G+7DtJbZhbNZ4vT1rzGGrax2/PtgHhK8MIaP7+1TSP
L0XekdjPsfN7PTnI6QcurObHTYOi8C4WQHe+Ypd3dMmN6GMifg8+qpZoJcsCvq/xEbAST+O62KIOWNquZky7Dp/Gln0b6aOLZIuQ
/rbQNAJjpeFttYJH8mKXsrzJAw5v5XLhh5XHgzgaQzCkzcgDJYYQNgIcXcCucQCHNTy9D6tmTdUsFVMSGgvo4Sx9KZAj/7QzutKq
BQbZJMuq6rZjGMFtfeFK2mqJLSNwGpeYwz5PshZdS/lpyVYYo8tXdxPI1SVqVyO1whdhhrngIBZpp/6nL1fXN9+zUk7NMKBVgWiT
ehbA3eQzx8wp5JMp/gXidab2rT4nzyjZ/PsCJiGc5syOk4uUj6dpqlsoQptENQOWN745/78AfHaxAiUrCZWit7OrU4B3XiyiVVpu
+R3O5CiZI5EY/YKnwFah7TFTBJyyh5O2YqMOg7BuPhhAku7VmJ0yfZEH91gp7wvSBajOaDctOahtpxOS2YoQlbr5MIG8a3/mmrVM
zvGqYaE324ZeYs2FpItqvolj8AB0EaH+rGT1WvjfQnMAyxKLT5sQ9V2C24c2I8w7S/CfESZxpMy2N1BbdEah4MDBTEvXpCtrzGrs
bl0FkVZTS70TWtYJM+9OCbUix0sp4Zb5KjgK0brxnR1gB2zRhetAke6QtJV91RTQOG2KdbEKoQG20JvPmAofknkOm+KYD72HzfXf
P4TP4rCu04lQt1U5his5DtOAZlfDw0b5Weo15fJ4YeP6zEpr8QniEc8Z46EmGU2iB1nY9/SXYtFq0h/NLvgOX9poR0w3wPc/91Si
GIr/w2vYJzkwBMiAiATAt+jIVBcYsyFRU4asiLO9EsB2sjRm+4Ow9DxSAunjyPEWv90Tcp5mJpulrk7s1aSQvUElelRE+x+r9HkF
iTB/rwoGSLsJK3Z5HMaT7JYVWmzwaIM7d2XVwirQP3YcxjUkyrmDDI1l9/ORMDgFpsMdYvnYywB+h4BzNcBNk3wLaN5DX18obaUG
/6SFXBO7nk2LyZsZGAjH5CJda1L8Tyk07f8Cp91WCZGv2EEYq8k0Dnv43J4eS2m0rg3Zgs93FAF9iuvVLdCfcWl3JSl1i1RykvwL
59JWPopp1WW6oxG1ypp4llaA034/FBzBLs/Hj+K1BsXv6arlLDBwXQ4HdpBkSzRMfXfrqUdcMMRSOt0/jJAA+Bj1No9hqgGN/Yr/
efUy7RFzwVWKfdgykspALb7mSRFBBQb9d2EW95e9JCM8cw25tm3JVGDpQggMZ5vYwcJADO7PZpoTdKQriOMxHcCD2Oef9XdXKGGD
lhZ17/n8Hch/6qK1MMZ/5zNxk7mvMWvkWw0i2PcKJbo1HyVySZOGGHXyh4pns5WS+hYCM6b7rd9fx/uzrQB3HYNHh0Ryfv5NJ9NS
ZcaHCUh1ll9K0CdFess0vq2uWIZoPqnl4Mku3VFVa2h8OKNuNkJ2zZOG5OSO39pQjKB5CDs9XJPhMTdkZEYklPjVk7wPdIK+4yKK
QrMIF3Qu8Gh1O3z+va9JgLpom56+VLehvgbFaz6Othi/yxTVejmPTN/962Pfze7vyzShgeJ+MS3+VJsd+KBtiFWasbo6EdKA0uvl
LyqghBt0nPpfPj649C4uy7i0G4cEYFvLSHDokK8QvhFZkFqsaBsLB0Nx/J/WSPJ+ZJuNHuUHuVYiDFPK6hw5giCT8huyrIrq3Ji3
8BoBkToZtqoMFQYEuR8cyfYsyKiMwiVShEbD3fOo8plVXc4RnC7yFBArscpGALvb9fu7wGyziecmuyi8gM2a7egYs4HKuasMPYY3
6gFwRGotUMe764z3+3aVzKpKcxC6qGXbBPxE9Qb0dVPKMm1jQnwxibIVUhkeAFZgnyQIuy2qlWaMcMYLM6YMzFOfWv9C0AesDhXh
fvyHzUI+Cve/TF8eBZ+HnN6ZBd4nsXRip2y0kF86geuJ4w/VVTDtnh7q5G6d36EpNsFEd3tBSFSm8cTSgP7MKyXs+y8cwAVWLSsj
sFmcc7kNr2gBh3dSuBSvGOyY7lfaUfQGoQdXhv1XEVZyKsCbDGAWD8ixV6GG5aKNdyNBH1rGLaeBjMaTHjVnbk5F5D1cDXPrnsEX
St9hUC80BsOO9J+4PdKiaLeUpnAUW2bogkhcF3zgvPC1WwgacEwx0qO/QOg1/TaWOJbWji20Dg/9dyikPdp8yXRvBqgWTeqzqrjo
2pL5eNMeoq8NP68Zfwmk1rR5YqDo6pdxcPmWZlWMCdpdLbzpKBQFF7SMYWFD1F/4nt6qDGV9pVvSyabazLPwXykfES6yYQ4IHVoS
dCkfG8opy2NRY+MH/N16KbJvZjV6Wnr/jfpBjWv9tMM/d6n9/JqYQz9zZPQr+XO/oywA+bmG/TRjbxVTKcf3QxCJsCNvAyQddlH9
cVz6/wWZs7vcIQ9bnd2IxAm8kzWRiyB4/DNkTA9erECgIBkwRNtfMNkacqnuAdckV1X/XECZzQk/MzRs1ygz0cmkiTqDrOpBZIJD
9HsRWKsFs324G4buSv92Bqzcyhg48Eomc2d7O7cvThLaCAJvf8vsCDJ2SWYsSIoJdDKRM4QqKCbyUq+xh6BNjFKNdfWNo3QhsXMn
dye4craXAj6q5mNy/WhcEqcu9LOURZq9Nip4F1oz2KMjbiHnwoY+4uhbp4rEComltorB6rb/bsUJUam53o7A+GTASFhotS5o+0ET
bNlH8iZXohsuDoZOfUGZRKA54Kq5UF25ACzYlpoYwgA8rl4X6K01vstGRI24cyPONF+jr4oqXXNGitkan0+bAfxSV0KQAvSI1TUj
m3OK6pURmafgfMQUjeJnB4BtKG7mdyV0wzD8tg3zvyrT2g49LYgZgxD4UDNsoWAeBxFxILxwZ5euSxv1Lcu9/G9/BMnlapYGYu81
Ahi7FoFaOCfjV609Flocu4XjIUAK3FQn6sbr5TmI8XCxxj0ERBcn1/dt9+nNSIMZ0AgtGqmy+577qD8GGzrmMTJUUUlzwBvnruDf
M/VTIMmD24yHS/9fodPbh+b2vr9PhgWKuIykozBWIWFBSr+l87c0nf8nw2dyg4kkTO5+y4/eBtLfihey/L+zyDUc1pZdVZUsDeem
VJv4GBq4bGgFCjuAb9vExy+h9/rBDh698CqUWY0ghm1pj8qKBp3HoIc8GpAnxM1RyjbS7vc5ZlIhoCBoICZNe7i2hFMGdQPlP9WE
L5sxxwNJ1ct7WBeeuT4+ZoS1MeHzV9/mrCDNr1cU0wDZ5Fa3ZTSpUIUC2guw68IAtrpyOUAtjuqFRwnc3xeP30N+/CwdrPrlZmVV
YPNd8L3e4QaNPLlUJkEGJErHKy//JEIuGdcOw3cFmzyfGmIgssF/E0+Uhl5WAk8NtFme5RycjMLf56n39TNVgWxG9jeA6e29KXv2
ki/UUVCnZG/hYhR+X53aN9XhPdaSktoMCDgtLO8rKFIRiWF45nkJpSX6g9IAfuVWZdaZqhbcIAp5Ie3ONjTNoLq2ql3Q0QfMKFPi
1KfQfyDTmxXEFhCs1dlHyMz4nw9y5JfRGP5Ec8Cr/jSctjGtMmBfW1eItL85W9V5NW52jw5dgpbJ9UCzi8jG23B++E183PiOSzBt
HOIEeoySxhq/LnlwNrXTiGJ4+PkT1Q8KmZQoHSDTpm/9P67rsfA5PISwyvzJhDbxXN0n2FzB9wsQzyqy3e04hvsns7EbYSkaNLTn
FG8KdXFvNxLXp5kd/anFw7PTDIPU7YhNlJFuEotkGuJit+FTUvP4paS680dp1uOwbnLp8sU75qQZlRCGNSkjGGOJYqfyVgyeq0FZ
Apz3zMNYO2f48sIZ8So+zFgMsH/BnxsA7GRBu7HicpzSaPUc5v744OyJtw4S2CzKrQLMFgXOx78v9Nw0QMdfixDGaJZpg5iqe5Cl
B+KSCpSOcPtdxR4q1hku1Mpt6RbbpZsGwzsZqRN4HHL6tzIQRavpWE3p81QWf8LaF4b6Brc0UBcNYhXOCP4wTVRARj9jlgJLa+VC
bZGxyFCZIJJjn/7mwlti84WDx4HGhbFp6EJbyYBByS4Ju5P81WrSIGHM6G6EZ6PCGFMpu3AgHzTE7YO/GGHNYfllhy+jzDKtelhz
QE+/kmJRY684zhXo98Mzsa4jVODlKTldp6wuz8WD0JH+OuIV36ZsU/BncnQujQ3q9hFGgD/3Urb1Snrnad0WumN9u3veN2lwkKos
8UkteUycy2uGvIlVrldYHbnRBeT9ZWoFVq2F7+pZlO/f+VilwepAjmpE/OSb/rCjikynnZYZIADHYQzTvm8bjQLQylNdGsA2ro+F
u4RN/FK2O2D/l5iWjJIgzzdsYOmL0SpOKCF8m+oAdZg/nLIvuLgkF9b3zpMM4U40+j7d5WVMUekJoD8abYGNbxowQIyFee6+E+Js
1nJJZ+KMXX0vetVat1fWdPTPHDl3i4pxN22I6LUmflh639doDRBRQ2vxE02aTZhBgIM2FZutzcb/W4OSPk45aiX8fAbyNngJ+CeD
rf+y5pe7CCkjQmLbbp/Y5Ir7SOsfiIbMYpgis4aQagHf1c6qp/O70eXWJdFJX0DujRr7XK5u2SZpzfEGP77GWNyFWTLBfVYfFluK
VfKsvMuExHIbl0apuHO6+5q10kJ2eZURc8/C20HZpwdbQRuhfzMCvrE7AJ5+nUZYZeLSEgQcwFhzQ/rd5rt6zNo7cicAIeKcdxl6
vFiXUFLfO1Cbd8HiwnzbZCHhmgyPq568DBbpUb5O7ysKwY1MhkE7kcikbuzAfAofK56rYtx92ckmXRGqCxb+mWxn4iFVufnF0uvH
1gafe+nbbb4fof0fO2CE7gMKVXBLd98WspJAcEBdfAHjeJp6tOk7v8JwBMT9tQQBZmsXjUAoCCmHpsq8TeStjR57VxYkvHMj/jdQ
6vnIPufMQkMPcjxzzxwPo4KZkL7baDda8BYiYN7pizk3E55FD4DsUXzlgE2eMZfboLZsyiI2dt79H9rNjqOHCw7isYEzzOrpe4CM
JtIjsigFZ/7bWgvivkzo7Lf1lpys6ynMH7BwdwHo0nsc1FmqWZ5ogPowvQGGjkITWd4sq0VVPqOaAs0ipVIxY8OISHHcjLATgSW/
AKXMXqge7QFIsGQf7iy8dlA4a+qUFNmJmD26qfrt4G9g9wJcdCAc3xUlYdi9bK6llxfgm9fEV270OKNnjtp+XHgCgv7+7+J2dnm2
dCMSFL759txkMk4vF+KkxeyFVvjgmZif6R0CIEmMW800jfcWH4SBI6roXn7Pa19nsULWkRcNuqB3ifmUpuymp+Vw1eYida/ZI35U
+ElN9k+jSzkhArYUHUMK+JFLOxA4Iiu0rcQ2AN4X+n9M2JL+TTFQt6rwE12boiW8+OgORHNGSYd7czrsyBeZ4Bbdq7cKCDzFQ3/P
6r2qvrAyCCc6zPMgcwDwIliHscsbKKAuEbpN4rEg4vGtVUJoaa/KZpI+fCzwr3v/Rfogioe4jhuNeiSSg5Rc9Z/1mLq5R6UYbDGx
q5GPJJ6KyuQJR7FJ/ajabbQV4ObdjdCw/iQ56gA7dEaSNhP0DeWrltj6x1lNw+x6ucj0iXaf8aYZ43LnRiewTxd9AomnWhBukda5
hRqGtYm1bOiSTkAlw8MeR1K5WGRmoW8GDHYKUulgVz2hcXMJm7CO80qJWDM7VefEqdPVoLycg0hnBVenuI2bc2sPTGL5dayjA/nk
Lf7RZ/RwcJ+rstMhZKWzMKbQQ1HLh8f4UFQZy3E5LCJqhr+m4TcCaOTBivA2I6eGwoI7rjVoGoezzhPlMvi1bMheBvUnvgAG95m9
CLF2woJi70iuU6g/IGq1RCFmUG5mpW2VemjeaBHWWuYcd/dhraP7aHyBNk2mXaFtC6vkZqSuGaErsf5712k6G5HIreMuSPhEmDfN
fPC+aec8vB+mmyfFp8wr755nplX5qAPj7kte56TKnYV0qAf6c5ftnRhgPnQqg2L1rDSbqr1252OeJA7VJkKWir5CGaobVlFxz/Sq
dhkFfaaa97Z7OyLXCmpopeKaQamZWUib0laaO8F6NBlIK4ppfDxVLLxthVkywoZtnFs5HjXJk7eR5dtv0YG5h6g/l/0hxt39I4Cz
34HInLv15uipf3SdBmQPQ5+jiTAjLPDnAEGDvrmS8Z7ca8TnRWy53+//yfK9ZcVr3Fe+iaxoFNI1e98xtO0qZoBiuW1p+5NVLuKR
Vnn9wPQt4hioaY5yMmSm8TjZ1XE8E/o+q2gqbBAnWO86sX+XHi8qlHDX2eYZYLHQ0BDFmyKzM14kMKJskaCBoVAXJrm5OO8yA8nD
eK+5jkRWfghCg/4foxjTXUcdILGGBOT71afvbIMc9mL/bgRqXHvlNc4cgXl70buquCyQZfzJg8smWMFBIL1aQ68iRRsgoEnU9adV
MTmmRphvJbnCLbRt0E3H2E305jeuLnEqcRAMTQIszgK5leZj+0UCjhT7Aua8DRVC99/hP6Era6mnLX1/p1a70WeL3IlDIMaX1j17
lEtU64uxq7vU2XVpatAEp0qkcBbZOInLJAY3MC36QZC5D/5h/y8sOuUvSXJmQi0rTKvjNxxqt1okGCCQVxl5La3+8Sbmc9LITRcs
Kiy1csA+jrhiuvTXIrz6S19+iFr8TbrIHaRUP4pTh/IbChiBnfP2sV2NQafb/Uswd7NvEvR/F34C67hTyEkA88aRSsAfj67cSRnH
OOXl18RfXzuwserGVD0znUSZUKthydo/iEk1pYP0zE9Ie0mwFwA+bcBcX16W9Frd5hUUaR7ymQLr0MAVlEj9zr1c9Fsnhp8HmveU
yL4EYIzuNIA6Av9Jh1XjzXjhySdxfKlauoyhz/UxwOu9J2n5a6rpmrkXrC2YBkhH+NWuTllznh8nOnmahmNas8xxc3ZgLFPZS9dP
CoLECNsDfNQs8p4q1m8vTCe8KzeWRsKEjh29dwDZiOVn4+nvIiVAP5R4xvo213MLmBV7wPCQ+jisf+742De6DX7b5EGhNkKtT9oR
Jg8nX6ayu0yiGoDAJl61cMZnLC0SYvWr2ru7vKARshmo0wX3wVxDCvR7cB0N6l0QbGpzoxeIiB0dFgDQrcpBVTFJEGI8otPYt9/B
0d5vqYiMrCbhBY5pbQbpuetKt85CGCqLR10DOlRiYh1D4mPP3aCL+s4vDoGzSu3pTRU5gS8o3pETeqpvT23UF2vahoeUKf2oJkd7
JvtDcB59RvM/8KiwL6RN9pagNmuMt1WlgWHdfGsRucD6YC0SRUKN7WHzwREWqcwS90+qd4hjqDtJNWpstX5qTQ1YUzeRfvW3Ctzo
GaR27fI0Fqz3zoLnetAh+67bQr5Z8LhEBsTFftFq3ABAQeebty5oe8GZM/kSS/A0b8Graqh/AIq7NtWDO7IAO+cxvxu7qgsVQ2o2
cH/7L/cxzQ6ZOnJd6D9O75uVFswAvgOpt1A1EWSb2vSAWi90AFBD7gJrAxZK83RfB+VEOtI5E7NXwdRAoVDfBFVq9H5wb/aYTySZ
Ysee/LrFhxnSftdRTaSE8yWLx91Z2CCUNRSEBxkBvMi5RBzTzU38G8QJBKoja6wn2yvpRpjIB27AMD9g4HBFmYWdnntE2gXKjduu
LHJOCqosOHZQIi8HOawBVYHj0Yyd5lCFvj/w465qvFZj7ARIEAy84kYczZZ+RmoCnWA9kVVaxZhRVYe+TcLVhkjC+6jVNZTDJrh6
Pj+XHDwfA/2UOJdMz3cPu8Frq/8ACcCXngyLpujLix5t3q7aC5GFbHhkgGEG/auy/4o2oDsE4szQJpDjcLDc7Uf0tAcurBQBxyRJ
76AGTynAsBlsrvbWFFeNj0MHoJF6ax3GPn5Fjgrv7O5JAgsu3W3mMuCaTYAWseYciChHqX9T4e0qBVFsT7tLuWqBHZLSOsY6vW5/
fSqNVV0XG/4Yu8MhBpLrBIibifzepA2F6rINXPfNC9G4K3cl6yIoqLUU7k1+nm4oU4XpxnwezCtN+stMyB8OKWyFMj9Xqhyhj7An
WifzB1hPTTFXCGQs4T6ygj20ZAKZnYvB0Z41ItDO6gxGVfdPLZTpXKcOuisNN0gQGwFUbSd6/TmdHoDnA7mQR4LVsBbdKz8hqOY7
RWCY2rvf2hQU8zhc//YjytO9rIswYyJYjMaSPoUVJnWiATxj6tEoy7Abo+6Ff0hknYa9+v+dQdzeDv8kgDofVzmoRPTAzCLTCaQN
BLe0qB3q1bj9ndYbi0xgojYOrH7LQQNR56U2ds1IuHkEn1qCXf2XUkQvZ+BLuy9bWC/OKBdTMKJMNzQdoNwjcDAHSmt6Yrhr8f1a
Hni8XG6858Clp5BP8WV2kLwJKgB4it+LfGkdL4tsSXxcT+Y9QFAmXXeGV2XF6ePLAI7uT82bLnUJvnRDkn6Wm77yc8A9Oy1QrdsX
Yq5fEJ3I4GfKSW69SrSmvcfJ7tDPOgwsJ88t/9iwb6zR6+bd2mh/HXg0ckuxJCQMn/h6wbHhFWgY1COtbAStkgMo6MszODF1IoaT
oo4rku/hEYAdHLoNuVE7BZJYRLvIFd0RGogTf1jw1e0YETDVLv0VP3k++XBGJcgiLwNTnfvdWUuia8HdgAf/tudKeOzT2eJy/HGX
bAnyS286vjiuj1GpypIElGryRDx9Q6jl4F2z/uJeZfAcFFUnpRici2RLQaXiRFpSzekYAEpZCZCl+IdHyA7oPekcMp5nn9RImRmU
xGUbpZarKE0ppV+bU7yNvvXnderMIDg9mewmwIEgn7NZ6QXmbcHbacjX11ALgolFjl1v3SzP7PtGGPqiEXL2x9/13rMbKjLs2zBJ
sTU/6CcBFU3+yGX+r31QlAN/gGXR0MPwyPAJoco4pwR/rBZVNOxdlPMStOdFcOUX7JNDNRZHwGxyY2ZwcQHq1T42cFIWQVEGeqc+
qnXyqxz1feGvEVtcWp/Da43CtVaTRe+60exdHPMYU40Wnq2XHMq34KrzYjz38iIurqg7CheCr3Uc3kvoDAvs+n5w6+cNedjW6mvw
BuvHLe7G6kQEqKY2bn4emiIw2zliTfierEjpu9LEIxqZZP3PC+iPl2Sa2oKdigJcVqlw9/951OqF7X6QBeJ4VFIUtC+1qF6mtl0o
/4/+5wkcpwuenuBEB6DkipkjmjwhlgmgdH1CGELakypq7NA7eMjx2eB2tnS61bZaGv/7j3rn/HgOeHU9Q0/T0IfI7ttHK6pXzJyY
MqjSBqe3bcDyf0stiH9PpFrpwdNQM3PXUd1SPcCbapUqLiGiBDVGmd2qTnzHk+T4SVwvyQrBRVCN4dbAUwaG/VCkqxvhu6GlbwT4
bm15yleJNKLoAMH3KkTDu90/00jbKuEANLWmNvMdPVN1CvjVpzgCoQB4qYvJHlws3W9Oj6cu9xFTKqJhP58JafdABLFGHwHoCGG9
OoAOgjNQDfgkfXyn6sjPxgDi2cJGusR1cmK7vRUznsgqAdTndm9u/CA8YoAw8me1oWlO15EPp4KbQLxRWwiIFvcjBguQL4kvQu05
S2r6GOqiqddmmhDSxcj4xGZ/5DXDb4fYawTDIP2gam/Zy8oxhqwBuiqxm9eoIRmuEgkfa9qMnBhh10RdYnWPf7P9/mSjr/m1r1Mw
p7VI3T9+Ygdpld4F8D/tiW8iCIh+fAwpeWarfndJbG+CX5obJScv9b4z/Y0jQr8MFBFQdnYvZ/asNAhoQ5QRdzzuDMsm5+4hnQta
gVieDza1FP3ui+cux0+uWprmaBkQca5b7kfaTp/n2MdDKKR4/5xQT+E3+VhNPJ6ScKoC/qhwlTh0XfsfZcAyy96PpATmqAyai7J/
00u03LkyOwbsrokvyhdctdP0cYmRoxAGYq7CYcg7e6yPQGbp36ukvPCgkoSn3ca4vMGwH3E1hfOri4pTfZ2kU9+CXLvbHPsde/fg
3qmKFQ4ILjRs0lfR2lXfjcVTWXzeFjjDzoyBLDPfRJYjAOz/8aogJyinceyIsntJNZWL/zaa1eiMHzQcWE4cIOU+6NHSTiKrqypN
zhdDR1ZXoRRQdMk9xgeORCWKfn6VddxJ7MyUs+9s0UUVdxaHOS5/52O2E57dd/6yOTNDB1Zm8apP5+RYMF73zzcLhm0UREOyKZ8B
4BSeflytnHAKHadENlxHj8lnfqlH/4ab0EWeraB89wvS+9ugpeoNefWC00XuVeK1XkXuzLFg+VaXwwCq09oZpZrqZgIg3KLl+sbx
PtY3pCyQW5NqPTIAjX3D0ld/mxLL+g3TK2vw2vnPOLMMwWv+rabigQGewph/IjjVlO4p9DyHne4jByWJ9uqAmOsjHcMgx86dRqND
0JYcTHXOw+KwgvU90XgCJozRBWDNEcxW3H/8sL2W2Tcr2k1TLJXaUfJRtzy1r1pZsxlBxl7nidJj/1/U+ILy9053UDSkGFx6QhJf
+oHxb9/TxspyaJBf2s1/a1V3T8pEmjueL1txJdSz0KgRhoVXWz9ynY0jAetvOHIWAEEKMPF9/Ypb9KgwnKAontHzLjZfSJDsU915
yhkRQnBR2BVNOCZ+zvPplkjb4DTR0YT+duB37BTaXFyeuhdXNjAH6aQT/1E7bAAbVKmFAgg82MxaToLMRqTW6orEEuZHx1NNgFKk
G8z4J4J21mpdvp7L3MJ+0hdrvBZcnHAB6qPWdrvRAsvq/MTXc+xcPpDV3n+OBCJk/E4Oh5UagdyQ+Y0EL8NavTu8mteSCF2Uy+qK
2tyOx3h4j02d0ScM9mvuqcJXG2NyBjCfmKVr3R7XcVyrRY//fEM8VIM0aImjC4r5KmPOCxE14l19dJmhev/2bMyu6Mmda0v2ILTh
k1UkQZkqWgd/HHQOr9do68IjiNno6/tUTSIU7KFtDVe8gx9zuyhY/IstUMhf06YDgKSP47CxWPn+ZePb7lDoZp1X/d4z2wgviXQo
IFq4qcpiO5kLvqBGJM1r0aUDw77mQIuoHzvrVIOf8QhqR72eYee9c6cAqvN4PleXwXmAnMimaDVcpgC68DBMCy8PhEfyPa9D305E
oKHJI16MnpCNEU0iMbXBZN1t+hTcR72ofTMXhQAwssNznmrOxVIQ/9gSDFLa4xxdkWOO6pqUlrwJJ+O6BpHu/F6jhSMc9O4mnTU/
AYGMb70uwwm9LyxJaYar7PgJOLoyR3dQl8djZa59/uZb/w8+pizRHSZ623nimfI/akx44yfs8MBZyF2iZjFA4L6ImWQMZWwx2DcG
pFmuATUZlvkmHbJjv3ebkxJndXDtdUxHNC27n5FlNYTf+61X8NmDEsyZq2PZgIj3qvBGaPG0f38J+8YIKmwYjASLrDWUaBtcTv98
At8niRSlI/3rCg7OD6kpe1qm4hnwcU92zt7R71ouakehNbYU6CnAJYrU8e0I2dpfcJFYKlnJQOXI9yvvKpZrJjlCJDVGP8FqDH+Y
ti4RDWY0Sbz9mWf6KJPOG8p2IH+qMfAbk9y8Uy8Nd5ed5OKGQkBH+tkhUqo7QkU0b5D9StuYK69oLwl9GcfUkzZChxp6DTk4IPh0
R64FPKCX/o+HyPmuj8cQGv4rCBkbJm0pUfymeASL6X97dMl3L3lPjbZEzwYe2neca6lkTDAvk5TNhyhO028Z/b/Nz0DM9D3BTZm8
Zt0nuMKzLxhekVOZtGWcFACZ6Jaij+OGtFKzMVHdp4M+Zau9ZZLmdhW7TXvC5zwEzcyEPiCdDOLhb830dso3mObJ9DyOpHlfo+9d
XdJjHaxNbZVehGptW5EybsHW+b0pKlcf5/TRk5z0ieiDzYf3giKIVzDhXcfhMV3NM6RXK358rcYNpP4fMQA1uoo1I/8U2dhlo7AH
yOkPX8p5QkTmAsItET+7OnS+P+3Q4J2cZedpulQdSCRY67moI/qeArukNwGKo20xRCVUYSGmkCs5reh3wA21Z2swOnY5CVEW+Ptu
SOy6XaU+6Kz0rOcYfiHT/GQwkINrrjxrnw9/UwumCpOr6+P9u5PnzmxOyRNvapiJLEVhENHqqKFSvLr908npWg4y7oaDMKS6Cre1
hTvq1qcW6Oxq9t995rtv2QvYt++2YRV9Svh5zQzhSFiZzolQeP7iZPGRALaGs+fLdVGY7dyyJ8qINZk0alLQXGuaqcU9/173L8KW
BnDao7LSSdyBl0e12Hnm+D62w/7ZX7l29SKBvxX332xirzK/Oolq0BEPMKrXU/suBRhQhnvuawksd6mu/66pMntVj+Kvqi3B9uxe
cIHZmUvjYSa5sIYsC76+C2aDe6fNiD4ogefj8XrwVFdQikupPln34mBMot0OVPEx9y9X17/ofQF00wf1EY0m4Hcw/GUnhLSEMfsr
VupisuImaR7941wD0P/xiHBiZ02MzPwJdl/+wlVdDPdTNsxgdalfw1s/wz5PqYAtanE3+cGMnpDIAlR1/WVmtRGmo0BUOeoWPsTj
a6fg9TUsp4pm3u4tXYOegSExUgDTf4BPoRBquJBWVsrmFpv8XN+hgtJbCqrmXSC4tNnufUfZmRV3M2fU6qw0if3oZ//KZ+QshjET
KN5KJxvI/ZPs/5M/GEaOq8zgswA2LwdUU27XYNjsq4GS+mQAIA1204NnrfhYfjQz+JzrJWRvHwygG/aCZvPWKolGqOhFe4z+1Z5d
euijJZvtQlOpSUhdy/IROSeYELBwN4T3iTFSpX1xxSCEm0nzbO+/MCRksuxRZVcxdkNcfCkV2h13NWx3e2muswK6Qob1CHKxggXQ
L+0xzxqT3lyPQ+DLV1Vb5hAujqKfCzLsyqCU/evJMAzjVnLIROamQxtrFzgQT+y5qazp5RpAab2N90WzM7oPGU4ltXreOLPwwwdR
+6bmXJkO8M/FBEOxuFgvGqd8+kpmzeXRjBK/ZOESXqQ3l5nEYteDAJlI8HA5hwylOTGQPHQAADGvVdDdYAQxWBklhnnFF40a5tA8
wqzFM7pdp3KAPcEIvu0kT2UvO7cpZf+0yjofd407KkDz9G0EN27xN12IBgeEJXmJV3+wuHkWxJc/7cKjVzOKP/vp0w9+0I7GF7nN
D9boT/B/lyXhSKT83qypSSFudQTZC7GpP13nb5/z7d9QUEMIudN5bTCtJGcOVVYqHL0Dp6/iLYrUDt1TXWlcelEZ6+H+LcqwGgj3
/COSe2z8uskWFa4hYJ9LEKPdvbx198TUeaieu0aPsstNaE0XIUj2y8pr4xxeeEmUFfpnQoyxKO6qcd13gNaWdc39tU7sCvxYAT6+
b3gC84GhiNpw568K5eppXDkwG1EUt63fAsB95FTxQhKSPzMoKxg8bGxwypmF6lv5YfhbEWR6Kmxli8GYLy8kBPjEQKy9zscuFXlk
R040rFclTxv8OFzmfyLn+LKk4Gy+VG8LjukaephMrwRrbddj0jgf3Znx5ny1XwjC0grOkc86dZCyebsfIAQPAa0afvRN6TciDWse
qVgxEDGKmeCORukP7iuL4eS0ShpNj/QdL8Vyd1N+mTZm6Zw/4RBG6/HRoijBG6dMaRzTpEG/PS37Emo9paElfK6YTmh5pAMvx6zJ
lFrJi3ajVL8w160RO5VDcM1+xhZBmKimD+pxkazvOXUwJhRsilznmdzVA3VgaHrPoevXKsIUpli11ilsY04deAgM6mw6rmj9Eub4
Tk74PezBu3Rs24rbDGq8cCFtSnbOzXMuGg77sPTW3WLJynLHAbC9NwsC/uh6KU4wnFmchBgAQnu7I+RXIcpr+Swyq5W2m8f6q2El
BwMCt6hi43AKoDmYoe6/Le3UDiiBa6mUCyrFRBeftdcjjNaiq+ET+xC4oW4X2qdW04UkvUpDv45NzMhq+6bWiHpTo3cKNyMSWBNt
RfflQ3c8xX8N6dXugXmGXdkfisA0DE4jq7FpBHfy4isg1qeUpOGX+ufIJJ+nXCkB6vVGtcOgFZpxrOoM8aE3tPArKMs6b6Mr1+Si
DhPJEf2u3pc1Ak6+I+HXNqPoEa6DrXg0zZEzBYtT5SWMRppUUyMRsu6mp/0COXDr+w2MzXMoZwVYv4lnuJupF2lDfnAqYdq2B3fU
8yWtLVMbxBox96Z3qz3wrrkCwtcxl4s33uCKGEmqWM+CdtUiP6srIfZUhtMp2vRJajUqyPaw8xShgLx4DbVKu2z4lpvk5lLa4nmA
sxvdS1x6xAYa15MU6JBeAuY0GmmsmMqAVvL7puez0b7l5/zXhW9kDkrlkrQCvU6T1/LoxLPgvOiWERCRhTuGVUeuY/3klI4f5iOg
qeb6y8LrcowooUM/I2IQ1Uwhto69Xd75pLR5YCbJ4FEHELVu0uAWwU3PMpoGNcHiECz5hpj5C70fqLVN4r2J/NY0ftLXY3pARbPP
OS+6MKi2k4ery6eRvUdplPIBkPmE/Etq909UZnIh4hOUCvcwskuRy4HxbxXNXFwLN5CWN9ctGD+YP6XK92xEhTvl7rc2EZTkpTl/
suk9Oc+0OzMfSSO2Uuc4H42ACXzzfUdFCxRFLHRNk7VWpq+zBNy/Oqyhb/BXxlXNqABi8hrVjfzRzmkXbxnFrpQTP4zg/ZMxS+2+
xE4Bs8gbYBqnpy0SwhAykCMclyuRsPG5Uk3EvxeEHiHdU4bDhTKmUXjJY4LGQapahsSE5nsXgSJuZXEYYuRQtuejygZxPlx7eykn
WA8v/fJpoCbCjefSxDiPHZkDVu87M7lg2W5Y4/+e9id5A+3zjg9+4+Mu4UmBKeqjquZavTNyslcjnP4xUfAICxn4ohbwRvSlqoyr
IGQFRcy+0cgKFUQ3xKciTkS8guGbbK40r+SwzqhWUXto071St6DL4QUJm9nSov0AIVV+Q5mWF7b5g5OwyjbJ5w4mswefrcLnr9Rk
fYa2Bd/ZnuPYf/3t+Wuw7otRWPSBwKCEgJY48N8BQpwksQriyMiY21YgNSrvfuNGsqIJPARV3HZRs7hio7fi2x54qVV4Ef/8nNPR
0I3pNUzZUrFEB5dOqEsA2Cbbm1uqekAopbE4MGUeIhPViQFwQjBw7N+FTEQ9xM+eFbgx/EtUM91ZxTQLzDRLj7G5p8YTdMG0+Bu1
h2O5s6rM2NtHKkt9uWwtxDwU9vsupeHUMRaqaI4uiSiBMLNAi372epc/Mjvf9K9/yX5T7/0cgh7pAnLteXNTgfyAcs/6gTmuHiMl
iPHa2dWE75pff9xWlnCk/aTJW0giIz+EUtjcpZapfUk7ZeQzz2b+Fjs3vFfb+MN6yawGBWzTr83cGm1SyNAFaGbYCrYrX+e2U13e
cqQxJJxZoybtUaA9XapJUEr2gvSjFVilpzgJJ+swHzHTJ0+1UR3hGV0HANZ9BePsEz9GyQsCnbMBR2lnYD3dB3o5S+oFaoCkwvWs
AGgcQYEnCudK9c8pyvPIdJD0C9BwKjudxjasdGiqH158NnKk2QNQk0vWhxyE5fb51WDDfRbQPOPneSZqyGdlYiQPm1d/fGkbcPP0
z923+X/OOVnzJ5Xnm4tDdDnkQsh1zh/aAkNyAYTu+TLsWlVKvaSZgdBb09RT6+81GPIV4w1h0m4L/ywCF6v8QHPJE/ph36YqUTyt
LzkdhUXxw4slo4J5nCq2865VRhCRbAUWh3oo2b6u0YUTSzhTRbnhuEWGL1UjS3+2Dv5YO/ppGKl4T37DJgkcXVmUmuuUGjXDuo+N
J6VADdyn8gpKtDOf+UCnoRnlIOMH+iSvhdI5nnnZmXjQKCFe3s5AnoJ4Je4YpG0jGilZFSWAz/aSYp4eCXbir83nFdHg9dR+VrBx
p18U/GNPBcKGjM3dY3RkXoDGqY2d/5R4OuQ5huLS5s+8qPc3Ry7VvMjY/MqzhyLR6dhzpHlR2vxfBukCbZYNKxpjeVi74eBuO88P
iNCX+ataF/plWjOdmT/pIlIWYAJpI2aCpgGoAwGjgEoQriC9tBJ2j/mceDzkUK6yoccEtNqX6bLFEdr0jW2kjr7PqQ5iGQY4Vvhx
LEH/BlPmOMrzf1VgVR045bS72Ps6lTtLsW7J9Qp7tHE8Jdo/YSrLVul3vCzMJeU4R3OS/7MrTc6tvIbN40jZIxxVWM90KQclXO9I
JSATnQo8De9zEfrbHUPRF1d0SnjxiTDJt84YS289eMfZ7+fHgxVK8JCr7j7ni0K6ZhD2N9lgKpLNA5u/Kblk/u0m90XI0Jux+c0G
J/awfQTEWApvqdYcKc5EFrBO0mOuz4HWFUk4bGxTE+nzCrL4D2YaLuT445KRcyw2nZeAURfQglQnMBCsKtLAanpvcl3mx/xqhnl4
EIAaivnkTpsznafgsFFaR1Os0eZyZDx+PzgmaKrlUeWluv/z0nnJfnu1k2cxOtsmFlmSzA5VaunIYjXO1I6QrqCURQzKCTNWK4MF
5iAic1w4sbgQpcNGXDFah+ppop6VEI3MDoa3EbXf2DGQJD0JiKJAuSaIOxY0gLkeXZuxQNKK/V9lzJndSMTH1AMu5d5UbTujz5yU
EnRNOJSIjqwsijysPlXdyr2Ijs6AZi+L4d4/hiZAveA/L4fZUeCHRTPYqoHD0UYLTcngHea4koxWTN5QZQaLIiPalFdu3ocDmowq
QuN8H17K6bIxiuozSp37k7wJNOF8MG9DWZZSZewpAr679DKrVbf2S8o6bK5rWMzQE0YYMog9DpsE44ZXyFl2m88nIpimYUtyRqAM
IVLwK4XLek+ULduKC7RGC797RwE39cZcTv2Yvt2prIdOkdWozxrg23N8uOKmsBieEVrU4R/X70u1RZ2pZJFTNvxM4tQ9rVW8BRdU
dDJC9aoRKruEUWJ5kzrw4I90WhFhivanoTRSX6D1MxWpRAM2awQiSNcTAImsNOFOZ8ddfvBXDJlDs18e3hx98yme8row7XWAjVxl
rQIs5E6q1xpEBdJqONXJ6qJ2b9gSLZeJcw9UNZBy4gPL2YLhR5NvgTb8kkc6+8Z6K6AjJeoWT4qdl8NyD6mhuppsKBaiUwhd8xq3
1pulAmT55UxGHUima+ogkihetLDmEIBoD9NLZRsOxmjzzBVthy4OVTtAtXaldbMqp4EuBLygrp8R/BGCwPsi6Rs9idEPQpJHAlb7
yyT9acrVhMd6j6k12bXH8IWMoA1vPztjXTg10oNA9mAr3l5qVLqGNGIo9E1PlOEv1ggkpnUKgnl3IVFPtYooWNt0TYpho6PHaKMT
Mii7SmkibaSizsLPZW1O6YKUva9qBsIznKzWOLnBcBAtTv10MiogThClY7pbrnSC2BHizdTWPiNmK5GEwKDZsfO/PF1uzXgMw7aJ
LfrFB+ilDQEFhCVZbccHzqzdw2flX9/U1E8NRthj8xnQWmkQPc7SIMLUHhXJZfaNhK6wft+nVwvSsjiJVcLoEVAvMtuHuUlRMwp7
4lrnfEmMh9Hl967H1MOl7dR77ZENOUDhNEO2QRS/RxiVqOszdndZdMlOOlmJU6u36nCcCTPYLvuuZ/WsIV6VocqzulE3LNb9AUhJ
oSIiDQ4IM57reMoUHUvHgfBQRtzxRfMRRYmDf6TUzEhWebWCAFfbFyI4SotSYC3Da0aoDOiuzIPtCjVtCAsQk4JPprE9walYX7iz
e6oHaX3B3ku+8He33TUNtpczt5RyIZd0SRivCzHieMJkI27VU4A8kXnaif54MM0s6w3XrufopOPG1KIP+UpZjPmXRoTOVkXeOdhJ
LJoXONmsCR0mLn416+KB3+QMnQ61aYJ+UyY5L0f3hVtQ4OGw3yOvSHcuzYkyZRuj0QjZeiN1eYahswr3hYHaAeFZNoTnhcSDlcEc
7ZZZwbDGkIR1RQvHTrCaOvThAOJJ8oIU0SEK+4d15ewU+GwOlBkxMU1FnM7L7oj1F+cCpxUjUTr9pTVzkqiTUkm5BkLJVmbpuB71
oIWWPXEo8xKoIMfBwCB6RXZEyBwWuSv4K3gKjYYQNIKJx3zHGzIoclf52mlGq+MV9uZu0liwCOyfggBI8yNCwVYE+0sLdZ6WPAlh
SHXARZTS+YoveDefcCuGEqJE24YU/P5oy5CF8QW38TZtvmrx1A2+nkbZJw4aRmRvNoVSE1aBpYA7ScvkuZ1uH5o8XQQcs1d4fX6y
4zi/jnSorbrPUaPkdQyVDfwbQuLW/NL5v53HwjJ5K5/bjHzx3UNUYdbJgXZWbEVSBdjZ6rXG0nSoTHgrZL83ZSPszCroXb2jtvTG
xg4Fdp8TNtg/woo7R+WQB4IQcLIGoF2yd9w32x1HNj5XUfDEbIjqupR/4ZFC12sKY/lXKHoMRbgfJBPmc2opaTZTZI8wNeqHyY1v
QN0nlrn2YRrGcFAq+/nhUeGiWcJh+PyRfV39C+eIGWjeera92icsQiOm/RcwexltstUDJxiAdwhLOVGsAek9rw5K/hw1TlHNhXyL
t6vGuMCstKIraG8Qsh7eltlSyjjeUi8H2QyfNKBONTNftnsW6iFD3Fuvd1XuYlXQ4ADLp4p5PjrVi/SRET+RUfQx2JNEOjKqaWZs
BP0dX+++pUJklGdNeCt56VmKEd5+gGwwzb9t2bEa5IvZ/MHftt+lxc1Q8DMud21VTq2r+GEhIq6eAhMiHQPJDJ5/oY7mA2CZICY9
dYO6d8XlakXC7lNFER0+FDaLlcUNlY9mtjla/XaecLvagAPPWLOYhX09+7hM5F6VhcAjPB5T48urbaZHwmiWkrabPzFAQGysz7JK
ENhOtvpGKkm6GxlXHnNhnEkTuQMj85qk9cGdDu8sUnuBI9VOp7XfFw/Z/vDhoPHrWlfeMUEXBuu7d9e0Vp2h7KoEprYcfBL3vPoo
+UZ07nA2Re4qPSvz5OYZTuefftrFYunpJ/PR+BtFs90DhwYr/6oRhvgSel6XSBe9gB01+kZPpd1H1dr+E7QwmHtILVpNiBf1MHvr
gOVj3yEJIOKv6qaWK5XE/UtarmE5OH3v+g/yQiH2iPv+4EC4BBwyrs86T6CeBdpBYEQz49r9DQ4IBUNVR4jNwsxnc1+mPA+bBZhN
5qNE2HZDKu6vSogYuqauKLrH416Jvp7BdGyQv6ailu4Ka499+aRS/+/w5VWqCQ3VCOyzBxJ15JpvhD1EKNY99HPDGUv4r1giiFwy
ft6nSOQ9R06INslQv1M6xnL0cKep8269BLoyclU2t7Ec8tSoT8cY9fGZAo0RibcM/erGmdAY2vs8BIiZruzoFM4yM8pekLRChRA6
z87TOr3IKWxGxO6Aj9Vo4suGFo0fpVO5F2nI3ECHcwAPmQJ5CkQ5vTJE+ky0iHVV0k5CKUdQI9CfgVYOEdKynJjsbRl7NRVuNFH8
mRJCq0/sUlcRlcAjseoUJ3DZcbzyxRfqk0RZKh3p8YEsNu5mndnufhpg4rvKTyynnrk4RDxSmEAUx72+yGbeuHIoJv1dXgq3WDXE
xo5iB+EcNzxO0QksiB/3O3m6olQqQf4Su0E6qwbLqL2eSWaSdDqlxhzrM+kRZLHTojBrlPoH+wKBvJDwGEBmw0U0ZI5S+pXIZ76b
NyBUYb3+W9xBHqJr02KPkqz6TScl+YTdYSA3jRm8hVqIO6NTMq7YiPDxC/o4Sksig0JS5Tvqq6sK1MfjiV1AiiOkv2VFDFKIP/9V
T4CcBs45dp0N6MNPactbNsZotDlCzVCSvvjwdG7ZHq/hIRLvZkVng/ygztj7Il23v4gaFMS9CjwILZgtxNHcZOFUknh0bxFbH9gE
cnf0JW4cL4vBDT+QLJB2gWlFdZpDzDiYYI6jsKoEcOrzBp7fYRFnknzUdc2A1GDCgayK7Rx6gMifeADVdQ2hDhFp2r6oXxbMlwlk
/vxXgoDxVBK43n0aFmqvCjDUKuXdYhj6T5Io7LQ4phJOzXzU0UZYuuZcIc1Ie6yBnJ4OGANW6QkeW0nlEIHMxZImdW1PMjkJWqi6
8FymQNKbici7Tg2CI57yujsn7kw98/8tqm4TvMYUF4TQv28U22w2Tqyw/YEjFWDg3OtzCkqSio2uWHCdbO+SavrP4xb+ABDBRHdJ
aulLjyZX0HELG7lTLZ1+iBiLjRG9R4ma/0L8BcaWRuIdxau5i+BW9VCAV0vmNj2ALSPgFZvyoXdaONEYTwYBYiIlqx6JoEDYR/Fw
M1G6qDJ3uofJsPS5RYaBQxYhrfwXE9l0VDGOnt8YwVv7FfYqpmev4uH5IfoJXOXQdEUQc6OuZL8srHUPJaZqbfdvkM9SpeH30cwI
kXfs1ye40Km27APExXA+76BTiyb68eis56cDV4sflO8/iIcMp/wrH9ieOTjyUU1j0h1PilyHhaNoS7IyCbaL4WBJl2elfPjA4jcO
OVpiKTmxtOMr9wvNph3UlJMcD9kRR7LRv+MOURlW1pFA+YYkLlmEw6SizJSJZQJPlE9Gt7ypiqoAFDr/VgJGb+Aqo8D/MJSxRHbk
OVrtWP4+A1Tw9p69f4k5VF6jrgDap561VSalNk7F+q8xHlsTZC7fGEHrmsqD0rfgRTkSCX8psA2rI2lEf34oUMWcMGDoaMitO+3t
47xl97ZopRJbe9TIQiwMIx6OqubwbD7fSgPt5Ba7+OfF3KKol3XdZ5TMQpX2n4zSh2084Vcp3YPV3RkKZAGcqiIzAg3rJpYePnYu
QArZew5F9A7hJe3ZK78fGJy+rdCs2CJLCFhaNKmEpljNtHICFiBaRolYT02yxWBGPfTkhypEQtRa19uVLb2qi4Qy12uc3FeFwFBR
FFyxAwdL4lQp4R8BjPS3ULjTu1xmSDjxOWZQXJ6N+2ZumR2qiI04dBvBC6M6J9gPfYP6IJM09HgRgn/Ciw1p/UWlCRNPI98EruQu
FrhikNSWrN1FKqbhKUutnyAbZNbOobXQTRpk80aAOY7w+eY1rQY4r1XybrH3YiBWW7MVTZYGgyCMvc++2eg/UDQP9IyZBAkXIl1C
SYJOLb6sHtbCoLrnC8sWCSaoYAAR+s5U4kDYjZOiFWR7xSCXL1Zz5jzWsJw/b5knTsVhtPGjBnnvASSKuWTFdCrQSOO6vO5fNaFh
nBQ7DqnndxhtIgf3ISTiPZAB3R0NPkB6A5kwxh4+yd2DOBSLHj+mU0PxO2ayZJr6Dks+XseWzQO+EyopqNtuR5Ew/aXOAeRRSkIH
W3L6UrYCjNdT5KEOpmBfv302sxp3A16YxE+EyUJ46twjst7wPgYSwVa/RPRZR3bm6rTj79c1Q2nccNiTdb3lrVaMLexHoBGzD/TT
52yKh+NJj1wjqYwD4ThNb8D1A6JOmB3IgPgsckrqOdcO6YrkXd+2KBKFYLgEy2ftmKgaJ9C3GhpkbLpSte9QkFdRo0YBmNhmdXXe
5U25Yl5n++avna8FB+1x0Emsu/jScMkP6/Wp+pd34mYNeIUM9VXlZwe1ZiM4BujhWHvdBN0ucf8w7qUaQi2+Z2t33lqee0DLli0y
OB7ydRm+pvIell8ZZXT6Q3iZhiBqxiR2GIpavbKQ32Znfy1aud9Rdt9xZQfiP+Xno2/A/CFtnMLMxgpUc2m/p11VG/O6+k5SAqEU
YnBoTMzrtw5kCfCMWJWgDmjeWWwrvij9sJ5mXPtyNaJpRueqmFHjnx32xUWV9l0LZIyzEAuIvbBoHDq7RktpzbDWCnjfzRPyrn2H
CZeAohqdtEFLEHCZIPj5trAMyz4biq6a0xvcoJmHb/2d1ZQC6B/Nzvz3H2P+g5tI6SJYJC7f7/t2i8zNdutdQ/u/HKa+x9iaoMDG
yPjScWaWd50kVvh9bAvEeOmkcwI8NVw1CVSqBjsax6sCkqng0fxvPY82tQejitkeRQGcSjWr4YyYho36Jfi9tUhuzQzPjrASUe3/
KaYZX/884YHuusr3UuqlRRR9qn3VIACNTkIy8EBVS+qlesZKmUcPElXIkBYg1S2vtK9HnBN5D61RkXCqjj4kKddmyARR7UALpCoJ
XMmEWtfmNxCJVLnJhmNIIv/KwiIWzmjV9EQIU5OH8LrxoAXreNVczxw/Cv9UGYtHOwfJzBcMof1Ss0h2uO8hEW28W4VRQ3ePuM72
g1XfmF5d6SGNssx4GY4aVQbzpjAMibH9lS8fPM+rCqJYtCMjhWzQYcbIS9RPhsg4Ae09fGEYwPodaZvAlgp+Km2TD9/R7A76/PiO
2LZFxwZYfpeX9EsilQs6uS1GFMe0vQlOCEJKjDSqetYJwwI/fppNmEIKaS76ywbuvnMaJ7WcQ4QcI53eT5yGEIA54MjUPkos9fBi
07xQkj46ucubBFEOZseI0rR+Iv0lONUGCStCzIk/RJG8InUV9daLo8yevvEchnHJ5D5ZEr2ARsKjKm3rIlqBct5zeyJzIDMoSRYM
wmgP7xs9kvFazTN9Rw3HJqIIs9ArGBR/OyWLHTX97VBfUKv3QJkv3rGgImjMQPoUAozSA6otW/68iNuqPdb5DHDAT62O/zwotuDA
TQSSGVGcoc7bHgSsn52C2kdyyyZc+yN8q+T3oqeBbvGCBFA7hrQPujBcN9fpAWXTc0YrIVnWDDDIrY7rdzWpnFlr30OHhm4ZbpZJ
EqQqDdD0OXhhklWFwDkHdGTmebcBPbZ8N5BaBzDNOwKv2SF7gxQGBw7KNJzpIsfAcghB8p0EBI9nrFnEmCNpyePR05KMLYQmBobu
X/Lrn5eq5/sNlZicDObZUwYC48QUlr38feEWlsVDaE4HxiBLlJ27EZEuTj7l4Yrv3jK8bh/gu+XmOJ0gJz3X3BVS1toMTa/RuEKw
2TkMzO20IMAAiQapBNFaWI07D0n7Nw1znk5LRC3/vxWfgW+Jz0QE1dBIRAy2UL1ny5jUxfPYH/26Tb3osjqH9ZvzX8YVxpxHjAwL
cmtgsOF/HH1oJyxrbs2aIwTUkut+2nO6v+2f44whcj2dPIv7IV/ibRBWlPqbWnpSuu2PZOEfzqctD7g6bhrWYJ8ZQBW2ZIjP7ulA
FJFY0Vd9A7ycKmJA0mC8ZBOB+6nzlvI1lpQBjdKu5owIqUEXYh02+BdSoUaRYoIEH8XjHJiisb3uAuYLPPkE1hGA2v47d7kHx1Pg
uNfY/jtMFnLZzMCvTOd6ZcnZ41WxC+jlK//6m/cbF+1shu3AWIjcsTD1+HltkfyQIKIPFlJPi6wd2Z7/sXsoSRMONPM8JPnXGc5l
94h9PzEwkBDVXyoVS7HgDrGjFdHluU5QnblRwet3ZUsdF//JJALkKFYv2sE3lOpgFZzzzwUY3pMNIuTiVygPgEDHbP4a1AQp7t94
DYra+XtbB5ygOBbTIiaU2K0EtMRQy0BSHbvUfVSTJEKA877K7NIQAyQxzrJo76dAoGXtc4ti49v04Q8OpMHmDv2R7GOpw1cfBHjN
ds2ssy8mEFb+7PAduAUVaQtK45coZmiLzKJdgiy075i6JZMyqXAmI4DuqmspuSUD//9zHR9l9wXcjcuWRijBqgLgowyxcSBigHIj
a4Gci5pHmKCyTFDrn8s06IHwUSmG1C1Wh716vxuhQC1jEMJ0/9iVWbTExHoI75Lug0yEP9DxbZuAZjhyn8RLc/t+uSCiim1l9LlF
G1ZjM0RdFQsbCeaejBRScSTUfop6yZgaCXS8pw9L2RM6xIEfKPXr9kHQMrBOmEwNWt0cu98I767avcUjrLDcShwnP+cclTltgvWV
DUUb0QsL9O0rluVS0puv+6C2X6UwDWVJ9xYdoOPmTGl9MvXRQ1W8DaIIV1ImW8slKzAkJEA8SMvd/W8+uKEMYP6tq+533OjPhVOL
y3q1QnrLGXtIQN9RoAqNLUdy/EURlPX+rEDg95whnpr82Y22RUBEO7JsmSiIG9WF7egMl0FmPARRNDIc6h+gPcpkzbFrQc8C35ml
D4z7IUNTfrYNBBOVnPME5I6PZ05hZhSGMsaVu3HzfDM9dSnzRB+5vY2oMVBKApQCxt0/1iMcefhcdusBeOVvLOIawT0AXYrMSOkM
lFxRjO4WA1bAiH0OUIZtVdcgBrFk2RMSYXndt/8h9t4rV7MXMcH+xwv0ojMadIGDymxx07FeRStSUrZAiqIRzXYKHHzHBnrKV/eY
VKfsOjuDTRl679tkS1DD1r1K12A8UB7+BAA4YWQyVKk8k5nYfv5WG+MTCioWbM8U7ieGYaukYFPWefmKD6cE1m1WV/swRT6cWiua
of1V0s275WtZP3xAcRS8qwjwonOb0Yyt7CGEJr/fdvykbI1nJcMDxtVHxHnxBvn8XlizldDHfSrDvFHUU+ylKKlBGPEqrG87wlhH
k/7dtPzPVwfNf9z/a5U93LWlDoDY/ga/7z93wp5GOUzkFwp3YJNnm3wgFRQcDPYDnQd2pprCg6HrMER6mjlVImolUpf0/IwjJ+fN
/Y8xf+qOUC9O2ly4KtjZiKdSAX38SxsFZbkHm18yy0d0zz2gVUVnXrSvTcuEL/C1Gr8O4g5JQByLSo5oKnExYmA+sh79aZ41ICjD
noHZyxfSFvEwRNxad7iupUZ7MDGZglZ2xtjL6RrvSMm9X5fA4j+lPkcCj6jYNkZaDSrEVjYEvK9RuDNSI2My0x3xaKAn1Gif1+ka
ajuNZLrwmVKwogn2xFH35hoGSQYYCd8Cozm1X/wAK2sA0cQ1SdQ/sw5ha3ai7KEpS1XHA1BZfITvG9U025SC+ZpVxTe+dbrNNMUf
n7/wa61WQ/BxldauXz/w75nShWuzJy0REnO58N3KAj0IYhRZvhwrRUdeIOMpEYqEBk33oakK5k79bUZP2u83pskrYv4n6KiAxCdG
agGrXwb70Xmtkjj6anSypUkQFwGz40qBEpmA7or2GGY21+t+raZUjMtITD0GFpIgbiKfSCYQf2IEOt88RjrdbTQYJoCed+0CngZH
tlcAwtjdl0VL/vOdLnRqQ4S8ZMLkNTMDnnZP0L7OcVezSW9OrECGzc1wkJk0/6pqqXkphqEl8/7Fp1iD1S/byLMJyRcbZfJaHF9u
nHDVvVXUwN7ibeTInOf+CUWOVvGgXeYi6/SdC4C42CIDfniG4rXTi7ysD4zrRINDLf12d8gqcT5sMOHSNqfSzDkWzBDw5tdKmo4w
hsFoInn6ithYT9uCh48oENmhSMB5+qYtzwiEOn0+HAtp9S1k05gJOfhiUS1iULo0AXN59/UZC6Qtio3U8isUHnkt0WyPt0F3v3tT
8K0AgcaFHydORRsZ5pbAua/BGPVy03u+qTyTK7MWDOCx1hPMA7Sc8Ns7le/NPrPTB5BqrrSIpNfud2C7tGvYvRrCfCadxvALff1P
tyV6rABhEW5E37Pio/R+yDyh3heMDFCIjgAvWZ9R3na5kvP3k2jyAg8zz/tv68y/X6322SCNhlWPv73E2jIFLBIL37DzUZSrPSpa
zLeUv1+czCyFPuhz9/xWW3Xf9vGBARVAQATcTnWxVmjrPM5GezMceZNlEV4WH0n+rMpAah7ANl6pfhejk74dfeknytNGJqzLwbqi
WEWS2yHWLEaO75dqqQ3fU9a3EtMjeh4gZnfyuFA0qxXMGkLjsa1krSM9BRG4DhrFVrwaA6JmVSeEpv9fbdRfPTab1Fb/OBaxb9+c
t+k6izhipEcqWCABL3zuM4G77yZ+VrriB3Jct8psD95idmt7Qfn0ovEPfk98uEl/o1uUo9uXrdPKIpO4II8aiFldUXEat29M3o2F
0xzeWz/pJqxQaZAdqiMmOMm1MmSvLeQBwPfQ4UAAiHPaQU7p301NmXRwDGipg3tK6AGqoXaEbO2XjiQsPQ1iskXt99AtsrRqY6o5
QtHmn/gmysb1I6zlqSovOgOTJx3wJKepLAkW4AH8n/AetEHABay3rYCm427ZfOaAbkAMwo/tkSmq7eZSWUzq6wV0GqPachjJW6Ko
hisxyZEbEY4osC/DQ2joFCHRRRSRajCaVm67eAPOo/2LY/V0lPtXvoFrp7ajb0NR6uC9fyulOBPq8tzfNGbiKAkUtDqn5tf+MCzD
aGoBb+lfTVmX93v27TMdneLkn3ZbMi5FncuI2CL7qU5G0GAbtU9j0OzFB2vuuAi9OCF47y7TefH3of4iYAU0cU3HjBcWjdQ6Nsyv
qRc7eeItfdMyy8HFlC5g+ohWuyKOyVWnU5sEzFxRO6actsLHnyGnxY+S7PHYcGgexY7sLpKvGRGT6uptYhQgOTrth2p1/0cpyHW1
mlAkFXkFJiqTWO/pA2Zdo/UelIDCeSoVklssFqrOoJnEhXGHSa34cw7fdeaXv5W67v1eh0ClwUJk7tRg6yZYZn2ZS/MlO/J2HANG
37KozhCd9PhdCA7e7m+kvOJ/86jcWguMFJj15WdUYSGlGtbyaq+usBCqKihf9zLt6hJ1pL1EZ/LJY/nK0hoRx9OEsd+84tf3GA3x
Jupf01L+G8wvxlGOOZ4iBsUBZY9z4+dnFPYHjNQGarHk0S9sDTu9/HbMJcXkpHgVEEEUo/lSonyYkgX8FDjP6o+1X97m09e77URK
LuI3EpzaLfTXBAIH0Dg8vEsUNKZwaoyeN6iiUR9mOYlvFMbpn4sYKOWDUyOy8Uhdn+ET1GIPjP0Jl7TRUVCe51UbtPwxkimgJUbG
S8K9PwKH2axENQxaR9c66jg5aK2N2+eEDVXe6/pNv6mBUAr7zYJwC3fsc9XmlroCOE8F4asnOvuPiDxg3YWn4kgFyg3aF6AmDeG2
dNTAbSffTrnL6inLnMuGMOA/fo+Rr7b6G3zzKVMiQrRg8hzc/Q6V7ki+6QHDBtXyZ8WdTqPQm7NlHPl9segfmBGOO5Y5DrslwfdI
sIE2bS372URlcfnYNji3kB9i5lbHpxGZHUS+tgPwijFPll9JuniJOQf96lSwhqIOi8aWozLz8BUnBrt4Y86cdH9jMM1lq3OtrDYT
yWELX7gaMLSv2iC9QRcsycMK7OF4M1GdVO/3pojszKZVc5fVgQ89S9SjdnXHROR008JfEF1Qx//Vk7aYEoV4HnDl/8GlryXJEWQR
ZeOwkagao4vDOYXUT/jPnfJgxvL8HwmIGVfJl/zLHTjZGF8AaGbnhgfJEqYfP6DHkGkinSp03Qbts2WCnPgBcVdUvaMnY+CJe+eW
p3e0VDsNHqqanSM+xaN8pRJelOkI+b1zpWfS4dgX9FlDaamYkRSSiWYwT5nADAcws93zbyiUCX2uBQcx+IGJ8IS028IkW62hy+38
cPkTvKtzdLL7n4EFVgJfas6rU/eHt0e0BtFe9uinAxOL1WzOIwQzQ/RhxI1Z+8nwXkhq/PzZgPmvtESvNY/LLmZVUlfcrg8SCYfB
421V9xN+ruEAgNGcQpEkC4RwAZHXiKYKkPXR51JdjchX4eaVyYO/qUq6p/6OoV4GnOT4Id20znRh0fzNFJcnKDytqfFKRTnATt50
E9spgJltXV+3E4IDEnbGFpG9pTs4ggXjJsWzwQyDstSJHKxZQxZhivc7rMQRSlVR+EXvK2LL2G4kEJNE/cUj2Pqybc5HJOOg9tj2
madrZ4a15866tsD39vINh68ciy9H2qRe3a6xJotiLHW9RP/w8HpjLVuinAoz8EStsjBB1Muced7ZyrX+cKEz49tVZIJAihw7+DlW
YHE/uCNvkuG/odeT1v5nEoBQvUyvSJNOt8R5EayX0jn5zfhc93mFxOjhZ2NeoykhWgUlUP3aj5ow0FFhNZXRwWvGkeO0k4ZDiV1W
0/WX7O/wgPi/vFqqDtSl4AQjrNXsn11HecO+0Oxg+ETPnS6So3VPnjjN347Qp8IQUEmuymGPIMv91WHYROYK3W2SpIzOfkIMltLL
1JidIwsCu2VNn2M8CXpua4n+Xr39i0mjzj3CSYzFirqX8VYb/Wr+DmjbLR58JxhP54Jfq4mExdpWPf/KBHDHKIciVSMgeGbuoI2O
3nx4r9gN3e56gqnJYSMUB0TxQYihB3HTdvuE5GOUf49d5cpqJrw7JA6qqzTXZd4WaRnmmD3rXSNggXBSv5iU3ccRzknuf8169erI
rHTbDV80UDYAHLjpJt64hDTjIze356b+7Epgfd1hrAyAC+mR1Chl5hV5bPk+Z2eSBtNswRLFYZtx2W5ZYBZekxcQWf87ukJGqlnX
Xf/EAG0EtOn8VBAKS394tPDS94X6gIgtmY5/Ko2LSga2tmTYAcNcRRgKw/ed9jbR0x4WkZ6P0ypKWT3s2ZZwQhlu75ClkD5H4jQr
BQ84nj1CTZE+F8q9qQeJhjECLspPLQPGX1/GV/TWKqVmPiM4qxnhjKFO1jPr+jRGT9l8afhqMJhEQW+mcjsxiOJ65ATu3eppr85S
hvPRqZn3JZi+cviuyqydrhcE1C4WXqMqtc4dENRk269gPgTN5wFg9PZE8WecPw63jpJOtkQw4bRX42a8RYK0E0COnvGDALyWD5wS
BIbGCkMw5+rRYFVlRpny4ObycEf8ltGdSNT4bLiZoohSU3gcuyNoFOybPEFlVfv+HHYHy4TG7eJGaIfqm9EiOMWBRcNQdJ3Sq0cT
BEPNQBcOvVuiHcMCKh8d/g8oD7NL77sVHs+1BtweAlFtdMS32QC5WC3kQxh+UEMGOQoIUf0RCiEMobrQG28rifbmLdvWzVKFr+f1
NNgvwc8P1ZljreBVghOzPB9v7GBUBntMwA3XTF5a7EqzXsyYdZ21HYyD7LslK4J+10Djm/ahXB2LtaGAFNFHdF0UG9Ddi4L5hcjh
HyhJYSKbDAtSe72h2ArVyp1PJh/CFtZccUEBE5IcEDUgqrXghCiF8dYbtEuiU7N8Nr1O4PSVJOEWSpMP0x4sSyBPUV4dkXtcA6bt
ryCcw9qaBKw8nSeu7Ua76aA6lhTRU4OOmaq8pmNEhgzQeAl/GQsbCBf9jHO1G+ODMhDj5DZd1vKiMDBgeW6IzxGCdOn+XyFD9T6S
oCjk+gAbtulPtE79rp1zHfbx2JQ+2pIgnicMF63XM+OMr9QAs9UR2D1VYlCHOD9s8ZjTPecSYxgVIr4CI0cuhmlsj8w3rvlFDx+y
RhkiuBGkUCvgwc58v1rHrhxuKz15fi5rs5nkoCsYf0Hne9tPApbTTPnsyeLo/DebzK6pc6v/RygwbhuXtwLNjZwxPwWzJJVhyNBP
sQXDi3zSJ7BJFLdHl5XeBrbGHoJTsIq/WTKuVqfd0ePHsCBHog6z6oh07P7d0BLlfL9vmI65OFE0KieKA5/2LrNfuKz0OIVEsNW0
nOsr9wqOyiZ0ts5KvWvxE8/CM6FFbaZwjxfuitcs1oGPXnjCSCOkNb/R0+k6OSS9kyfhdpeS4l8pSwvNSNot4hPeoHqRaBHn6GUe
1hlkWURG9RZkQua8ZHC7kr8+NG7ocgNqaE2ZeXbxFnJ2Zbn71X51P0hJ1t6Kpa/62iWkVdzgy/BrhpwKMYKXiFEzs7+FjJmuJqLU
AsJiCNQtxbfuAb/jByFhx72/NhabSAXRoObEYw/Q8WSMVXJahgcWbP9uCW8o5G1yxUpTxCvXSNaYYDG4M8EbYgEpEJ+yor7Z3rB3
VzDCbrZwnybBL8k47m9JJWH5k3slTpvyEEWqcgGUFi5pQOqlkThVr9os3b7meuDxkNothrPBfmcsYed/gXHEZC4lj3dlOH6dxMmG
Jfkdf0rDbONTXfKs5MesWsBQ5NFPF8SfyArZw+E11durrVUGmvbfm1E5rzfMjASpkrsZfTXm1IWJtAPBBzJf6yYmwF2YCGeU7oDg
rDKBrWNHsA+HcrFwCMIygvhK1M/M6AE2HGzqeg2yg2fEy/r8FuUZgUok4rPs/uhmPyF/KByjqiCocmrkcrD0PpRTGUeHR8x6a8KH
h/IXm3RplaIGp85/E+EUmSsFoLs1FobE8I6sUvqBWwN5qphv39NCrxGxTel62a0R/XgHgkM7eEXkr0s49ZFreJRlzQIHjc9LPpiX
i5Ob8nXz5CZMdLs8wRmGgbdaol0cqwt4XT7zAuJty1BpaeummtN8s/G/eNsQKPBfh8KGgkR0JN4l9BjhZ9Y9vAr/NE1XZHrZ122h
INQjsIAJXcNOTjK+I1emXsD8GU1en7Xp3p7ur9Bkc0Ak3rgWWCL8uWx4MeUA5CaPepl0hwVj075lhr7wfOFQBjhRoy49iAsf7dTR
r+gsYIoKGCmr64jkCc7lxkJC+YNABAnm2+iXtiuSPW75xMhakO/pLGAKlNtkaXFg5kh2u5j8aVUDhjwujkqq3tT7Vxb7jkQwNgMm
GY4Jwuw8zti7hGOlzgrm9ykoeyeOS8wETRoRE5Eyf6YUXmdo5qadMw14XkgAZ/58fjDW0Jcq1/PG5NfTFMuMazvxA/tr0QJ6+YG4
cg/ltfTApQRZOQeq4YWucixblZAvp/xxJXL0V2TNfcyu5JlzKP0NicfruEZsIhZ0ib+P+/S0YY599PeVrFbp/aflXMpwFxYmdMf6
RE+1roty611+zC1Mkwh/0P59WXJS0617pY8/A7BcoQrvdm13u84pGCy1a1117MOFfexxdCQe8a3gvOpQpbvEolr7flJjwjPntTgT
ficdYsudiDs+4LD0G4Zmi7FW4YNN6oPusrPnCa7TA2afCfRkGdON5RiLwMkY1IGsLh8On9ufED7S87FAEIi5WhZPPl4f9XYbbHeO
Xnn+TOrgvRGb//eiZtUBjbkX+d0H7/B8GppyWNtjRFfay3oTj9plC2EU8kr5IF5Ka8yqlLSnP2C/jtxdy0J9U88mJ/meTPppC5YN
hkhbw9HBSJtLmhqatxI6WRPQ3swlMRGaL2YMXFzvuCfIglG4LeYKglpec339prZJZTZZ9vbSY5c/eV4ILgYND1X6EIs3Z9Bjn2wv
84m7FDbGKBaSKrWkNo0XUsVz6VjaeLsD0fYq9xIX6Cbztbpw83jwNSA1qiyn7pLFIzTlb6Cf3aVIuJR2NzwyQolLJM4iDzUwyaHR
vtvVuEf7oxZCuMzpYxQrC8wB1FZOtMpv8C6BB4ErND9RqBQ8uYONQOVDtv40UWku8s4DNON8lQIiDo8eP9v+Vp2/xSF6ksGO6O1z
3o/v7QS5VFvRx2kF9Eni264MIgIFcJqBeRAcm6hT+bmZv0DSY202JssN6/tQsqZbOz5KuaKBf3xzrhh6DsORbd7V32CjvYueoR66
nO/l/b4DEDyXzQub0d8qlHVIpK8NdKM1UF0QzYPqXKjpwBueHZBXDkRa/3ggreeuViZCyTjSy6JFD8uSSb1Tfpna/PQZ3Isk20S+
3Q6mxtaMhqOh0z1XQGSM5qLe4gR6QxxqKDAGsnIWMCJn2YpIk/ZrhxxZ99xbVUxDoXkBTgXui9hAtibutb8a0zWkIuAMaV3HgfhY
RflRwyG3Wb0lbUwBxI7nVRdtc349STBjPjU5fn/G4E8i9+yao7wlN3iQKsqzoTWJHSGPL/pIbTZ4/Ya7qdCU3VdProOJYkeQb/el
7Gv0B896NbNr96kb5jLZ4ObeblUkv5ASI+Dr8Vxem4NgGT+9YizWExQV1xDtcpcqOP663BohWauVpTLfvbFlt+R/D2Rwro/ZlBN5
bSoDwyGJNuzJwd+HRvKrkozXHvsu8mSbfXZ9wc9uQ0f7gluJcD0Y/n25Sm3oTwfAVTWJcB9g5a2zbRwJmKt5OOM2oyoJ/uD1hQi9
DKfoU8oDmdW++CGkaCy34Z4vdPw2mVwuKkkwcQtydFZkSFeqRz84EnsjSqDuF+zDTOKjIZ2lHHD9DqDAfxitqC0FFWWKpzRx1Cva
q1B+SQ6UPxda48Hqw+YKlvMl1EVePYcW0d8e8fkUByRJiTXMiNvU8N1XIIlNiJeFeSnlXdMU+cut9p+6ASv+A3UdKXsOOeYG/Qhm
2YYu9UoPuNp8c3jHa3SxrDoel7uejh67TNekou946dQpJlCxU+wq9Qt2LBRTf4P+lbdQBgiqmLQbE7qOKji2rCL2NZ/hzDKETuQg
n4wmVmoBE/W0BfHh97SJdJi8BIb+61o75O1/9i0kpDccSUwf9/a/wxfc66bYoXiVA+tK5GLMl19WjSJKzGEvVO7GAOV9srXG5GHu
8qt1UwNGc7amT2X7OU+I7d03TfkFH3gHP+T/63OPlx962XOdbbyZGxPm5fPin1mfd5Jstj2vdlo/75O3w0iIn38W33HsWpHWvqfR
54tNQu+JpswHAETeQpV9SKX4qNywFPDBoKwM9GfOb93k1uqCSw8bwmpxtsEe8C+5WYD+v6qqhpJgu2yAiiHNCtrbUUWb8/BFn0nU
H0NDu6LUeIH9dpbyHkvZhrmctJEkF6/Je4QU6Chf53W0XNiUI2FIkJnP+sbNpoLmumQK0/r1xlIu71RNZYwBIwsie8iLycOr0Kh0
ncrbi4UoSBy7vsuAgnsKGgzcO1Rl7o8zyQ1kqUYKBljq/S7D5K4joAvwP0mJONvo7ZBAOOcdORj1xonsgFk2nzNGe9gUU80HHPQs
qNG15RVdaF8S1ocnshx/Avlc7BiFvj6YE2US41bqDT/wlY3jPQm/qWtyvnmsHNt0vy/3ZMuKrH7PuAZSLFp1fSRP1t2Fc7i3HU2R
MjZuPEMNk343aQM5lvhkFEmTOqN/NXbBxecy1uHXmYGjcU7KpGd6J1h/eCBjF+pEXwQcLuCBRq42xw9Ihnn1abZI7DEYEZqCwHbg
D5vI7lcM6a3PqFJrdHyrJzYYzDogVfEIlG7cHkZyQtCVzAZTldzTfq2khDQKO2QX4kyhrMZ9ND+/XkgoXp8tcrHEeFsS8acLySTc
Ax4MakAZn5NmvR2aiQlfnSUU8O9K5enkqBdJ0LAS+dBKBofpd+n6EfyAR4JCslwJD1gnG7Xpthzw/YnBwxggk8bg0e1dFbScZPor
jfday6OvM3mW6e29neWjtsk2jb7dsUO7ndU5adXobrHLRbtNKf0RMns9Or2B901KkjmOPUqubXqvXsAQ1ASep+Wj1bIfCizK8JSj
jRg14oAutv9mxjoePk5dZ/JzqPX9MIPbqzy54fR7le7Fh9peD8zJEv5bnUP0r8e5hIkPkR2avhXS6T+qaVsLFju3nAwKsWP+Pt5J
gUe2qnIIKIzAX2QbfjMf4LUR+fNbwFIg58EP6XTZ4cbtYjst/5aFB9b2vzTtzxSwWj8DC0zW66rpDeBtodTS516ZxwmbZzeeDSZT
j1j1Hs5p8ZSU0QllPmnItFj640EyIPGgAplw/MfKiAFECAE5xjFCftxANhfHQYxcJRUl9tvxQeb9ucs8YdrSWitOzCN3c2m+vJ/k
TNUdhx7CfNnMe22efIVynhGcKGv3PTzN/o8uxkvLgPokVuw2/1CTc8TL8O4Nl0WBlHqg38tR9pCM9z3EWT99AWEZ8BEaaqx39xAv
gJvgyBi31BypRT/fPYf9/MjHKlg3K31axSd8BeyBt2hvx1DmUv3TH0AzBv0Y39ImLI71e97rVkZOdJvW1ZMuiuwfLU34ifdIeJrs
kZR4i0xBsUuRkPCZhRysA2ngwKv/XvJc8+59P0VnQUQ/TuFUDcRtETI4uuZf3x8tNsuHFx6QwDmUHjiOVmEKKWOjUdSjNi3R4LR5
wgn1yocwQTnGADF/EUrE+aJMr7yxyhxl6QETkuANVE7PlGDfDKgQ49tXFUwyDJBCprHYVe/BGWEfV9xmpa1cKdY7i48BqlGfIyvq
LkH82JhBpkzGYfuhiwqx40VUUqSKJLfhU4zI06FAwJ6BaFg6VDzSCYcduJNZ1xPpnrWGP/ZYKG/FfUF3rvVCV03S6Ga3kgoJu+Pn
lPDhE1DH+9s2bNZ9gpUTFndK5NcePMBdD2kppeDHqfG1Vf3JdARB3li61K1BYq6X3uh5c35S2MpPCaixIrqohw54STc24NegAKYI
jzs9zIeFTGPT73vJbHqXS5dK49Z8aYvQdS1Bn9JjxkelnaBFygJGrr+VYnk92/8D+f4nzPqzoqqWxF64C22Yn26j2FSHCorX0QKN
54vrlvmYWK76UCr1y/VjPiLxS4TEivIdLr8ZFOxcp85LF/yzJpR/usyRU8M9PrIeLhoTVDLubI1+z5PkKix1C98b3vxpAx+53d6T
qcq/LSu3txMQmoVUr/9e+WUZ3YPiXLPQAJ28ma34ECkhkp7LF60A0eu9ntp149RimRRWgPmZYNyIvHn+6WBMpTv0kvt0DM9Eci15
N/YYu8HwQEJa6J6OomfZJAbVGibfarOuAEVxHihVuiNEpJEQVY8wjBBUwrftrOtLPv0CAvGtlYgMiCI1PYXf4mzq04qkbkzl5MQk
t3dvl9KQQu1aCth8DHKsc25DiYEQvxaL/oisanQ78dMv3a2wvJFDf2hhhKTQoE707K79s4ykux3ZO/VreBzC3p7hHt9guN9tsia2
0SCfgdhc2S4YEgvjAJ89n9CUXqW+Z8Zp8sl8ceFMVt2kIgCFx+LYxulv42n6tHipQOgQTbgwz0kt/rj5y5WgshZchO7lCzjow9+7
7Mom2AfFDQR6AA8sPNonJB4JH7KQ1e3DZtLvO2m+XO26nQQczJHB2oHvopSl1O1kzcetrbAAcSrmQHtce+YoZ8529zEVOJhFPLI/
K2YtIjR9lOL2dC9oNA8jb4753XMmL3FOlztMt9ErkLFdzYIK9S7bdrzn2daYBxxpwimbU4pwF9xj5fyBnDUBtL5EyWu6I6eKbRxZ
xlOCKXAJKYElp7zamBaehjtKTQzwbK6sDgHe7H/7VN1XtpUnSWT6UvOCeE5/PPc5Ehu7Qc9a+yMMu9qsNbkTknRct1QJYPbjheEl
HuBEjkxPmmK+MGkMMvSv5MI40ceCI3Q8xDFqey1XQNyzWihZeql0KX3J33X9W/TnTMXt4WLbasFWgH87eJgYBX8vB+f7jpuJUibH
CM53qDHq2u0n4bkMNasdqeQ19hYEyMY91UF/QozuHIMaS92GoAsRUwem1P7adYtNbEfqU+/m/aq2wfc231eQFajJqTCPxkZJo4/b
HNL/+sFc3EcvcGSuVVxquIl008Q9Vc1RFrYbf0HiAHZ0EIy+A7nnow7s7/oN9M9nR8Ufg5s2fXrND7LC5qVx5+mmp1H+LMUSALlk
oLR/dAXYxMeCDQE6Lf6KOV1OguukpVNCluLIZvdc132SMYvtvcfSkoMHfABGBVgyNI5JN9lFPTEmCFED4IqXmjN3GQEQOQ5/m31y
s9agI9rEXoi1f57g49PMXRQ4QF/kuRW6GsLUBENFctS1zWwAq9OeXhtxHyYHFTk25boybQqwSUYzngweq1VQTy1iijkvwowUC4vl
cX4lUVtGwMRTwCBUDVxdkeakF8CtT5DJAXjeKjRGPVObXpV1f/SNCyfi02YtwtEmptaUQHkgZ+J4Wa14UPib5vG6FS7+hLzXFUms
FA4j7ysJPRyGIH4n7v1gD4PBUSWPZKB5aQ1Cl3pTrAiDBV5zk2B6xTCJ83totMqsf0K0SC3LfUVv9X+tClV9qas8WCgMEoHuZYyw
xbeBszoGVeHrjKyKcw0B34dvhO1l451affDuI0GDA1UjvSP42oMK7S1I/Ue8/ku8eE3m7w1v4k+qeTmPkPZ4iRnRM4qZNcZSNOtM
+F/RTc2Nw47cz+qMGBvkNNZTwr1oULX5dFydbRVtW2/aEzX4RPrOuHW5M+WmYIqbltlWsm4KbIdotSdm+X1ZUqZKE9pcGdL96ykr
1EqX8Au0z2IvcSi6NAWorm1cBEt33ADvF3u0Y60wp8ViPqdxQvftTucQblEnvCN7ZedzjpG9E2Q6n5tXtXVV4Ty3HmZNVRWdWGMb
ZI4+xXwYNdPzu7bWdQdabQ2T21+XkTGS/ihnasCF7asvFE+8bUpS2/tDYnMZxf4rzPjelrADmUm1/VB+vEhc5lnnJZdB3xbZCSM8
nNqxmhVWdEs7B9lbVBazmkhwX15A5GrEt7pAOPQV3jn7dUZI9m/CfOZe3Fy+s6QgMPvrb1ipURU+uSJPy+3yU9R94BXN4+uQrK2z
Ua7myW9I99uPbtY6RYT6kFKZl8Zf66aN/UyRKkYsxziXJXkveuO5cubGgw/fR0LzHrE3N9zUP0LPgRzdkOeAx+IqzJYfNsqJvM+0
z/FYZCMBUl+3ZiIdm5moOYRQkUOWyAt5o56iGbAPXE0krOeVlewIk+0gMO/ajF7/WrCrRtEGS7grY2+ZtUslvJtZYnGqmjaTfsqs
v4yUrQtNrzadYCXK/Qk+gWiWgpK6fMLSFK9YVq5GqA0MKfBBy7lsSLH/J4p5Dc9AKIQHKPVF3iEU5hqNdyvEaVKSfo1hJIv6Iwpt
ulFrKQW2CTeRKz9exlCm9EKk6hdqIpKSCLNCWPWybKAI63lQvFNRIODRt3eqN1LnL/FRuMdVVhpjJMEKWxsq8YtkeeHGnR3/8f4i
99OgYNp4kXUy+odAHiXHLgfaC3Khzv7Wa5K7vrQHnEhP3Y5+lJLfxG91zW9yujENjuTPpHZvze2KdlnOr6cNpNRkP0U/QyIyaazi
3y6sq7SslhBcx4KonwGhpZ3rA9ClQzPPmvYGxX57A6bxbwD+l0tQE18spX27K56hWRdThBVeu46YPVs7vng6gEcdbZz0KiHvqnzy
EnI6+buvAmoC/hRa1WrouuEqBMGiUMNioMNw5flT+FnQu3mY4liTCbxcFR1iXImVRDEEYHQx18+TlhO0iCjNohB4DyLSTXIkNJIL
1lLVwHXX05RCga5o3SSH+wEaR06BRuRy+yKj+KYCC656W+j3E2iqQ+3otsdmBVxtE3ZNxfsHQoJE7UGci1iAI4p9UzsmNi29Lnt4
YTHgge49nJ05VKwYn+XU6o5+4Cn2F1UvEr0zgOTeadgabH7c7GqB11f6q7AT6pA+mnrgupqn3uxgbz2v9LoaLtCYvIAapc4ra67N
09rWHoqyGclGhMaDc0j4b/7IAa8t758RRB1WWGwyz6lB1BvOJhY17rlG52wJYfuT2fU1YnqwM67s+NS8aOzpRs/MBAapzdlw784L
zcIPPqOUCUUD83yl3j9QXBhCLKu3u9Uu530J5jCzAUTRhhTETYbcKnjfI4keHC5A+We/Fg7UchrXXVdUAM9FJLruWEZ4RnsQt1Wh
VDgeT1jaf+GOqWHOi8epcAi1MY4pL6ffNA0UmYCpF39ojAqnUKo25SEfYDEp2hn6CVtUaxy9/IFMEb312UR0VZAwp5IogBadjN9+
/e+o4rtlF9i0hv3/M+tCV1jHgKyU723EkVuPJ+J3QXiBiECP+WYvZI8syjguVv0RbLRIBrdgCSzOHynmR0RTe2QUKuPrYKYG0p92
vVJjT386nBKnL9mTkVkiDqlYiUdqw+MTJ93HOIFbhWjWydyf1Hkf2F2fa2+18dJy/350EGLjVTcZDxKg/bHPA4utdpaptqTYfojL
DmuM47+2Lbulb5fLFmfTxc48/tK1VvA1dReVven6nC0Mdc+0wLJuLEDnrFzVHdvRw62MgKm75TC+qThUSMIJZifWAZ6tJHVVW5Ef
EUyTtXWe/8zMSTHMu0zhCrxValrsy4bsKRJ3PrNutBpeNCK3px0eZjq1z2buHXQEqyUjP8IMojp36ZtP5LDdjPNjaFRhmk1iMrpQ
ox8YDW4C0lW3V8kjH8mjPh4I0cBkVnfyUBtwF+3FOetGD7pMfVpw2AIsPTzQkm2aSMYSkR8/xMzJVI4VfpnB+qeyOOXEZmLctjFC
y/ZInegtHhuVfvKO7O9XR4T18YrHFobyRkFmWCs6GL0PUEksySOwlNvb9z+kk9zNge5xQUZj7qOyHPwHKLySFFxM5Gynrbm++Pk6
7uKgA8DMJCA5CkH+oDiKXHtJqQqdse4KHRFzCcSd7FGPQFDvFR6QeOccYaE5QGMQ8YeZ08VuRe2xHUQHkZiL7m5oYYHMcuTP89l/
ian6/V4+rep7PhUsh8hpurrMFNC4scIASYGc/hbUUwuJR2ERpbbxKuuQNrqa5jx1RBmMdMnuw1CUeBsSyC+6IEM6YMffBwYZxRmb
+guW1heQcWu2scLmEcZ8gxfn760QyVRZfDZL7Itglkvyp/HI/vzlUiV2X6tM83LFTtO4AGYOuIOEoxxKVTtV1S6F/55Hgwja+op/
wLeKfGgYO20I6K71tzqigbd7C9xIf2325cePrF1Noc8k6BcvybelyiCnXrtmYHMVsHNuXYjsPyI2mgP2bMeIg/m8lPvpdEqbtPje
Nh7RVXfw85YSVo4ksrjmZkN6q2mnyjVA8qjQzbLoeY1UJCNbexAqTb9yugEGrlZ9KjllFo1GfD1FuAo/y/Ya7xI2YwT6nuLzCNiq
0JNPCUtHQbTfugNrAPP5X+3qaNrrYHgX8yBDJj/ZidV5/ZreuT5pSu0ndhp81x+gE9pXUp3DTMKLBgUu+xPFqJtXuNxcMh6yKRPW
/23+oI6GiP1UUpy/KUWZCrpA9j8IxOnieusbY+WYjcCpTlisY8wGBl2idPNkAZTrKU4Pj6v/QejBMfNenXGRsy1fFveTUuSKFCHE
rteK0+Y+uyfvoi7JSBcewWZ75XMvqOuRXaypj/v1TWVT9b2GpxQi+t3cv/E3r8Q433PMdd/qlZE4JUXh7m9pejEOpyTCdrmoO/VS
pHOA7fT2W7iU2AJYHCb7zdPR5MSMK682h66Hr+lVlCkyGsmhlvvR+95uuumfgEnRcsKJAiTzBGEpvRF6nHO9uctgN/+IazKDc9FF
lTK2I8b/GY+VT9IzKTHJZX17upEXthaTkh0yB5Ka1Lj4S7P6RWTTFkjOo078Zb/NtkAdJba8N9BXwXGzoNHbhiF3tLXy3k9RvdR5
ANYaPkg/M8eCBqtFjUH/snvWGlIGOnip3RzIWBYqOPRpuGZKUmShwlshD3GfDqvu4jHsMpAXeXIc6NWsVDrqKF5gpuaQnqmBNFBZ
p9h1w5Sd1TCPo9CBlauA9eWO2BxiN1yuAdGJvV/98vBiI0s9izVYnGIl2SCUBdkDxryIZQw8nw5bp5mcGQeNApCGmVqUf6TiLgCi
LYr+TD0ToyFM8ny7HmED4IGEISMqiiz1SppOPnkBvruoA3pevwNyMeY0UMYlAf1aPzP/alCarjDndHYjFoeHWyfMWzD57Zi7zJV1
4jx2rrGWikTPOVhiRUR5+TsIzglINrIbuen7CkP1jHELls4mnQ1+Grkq2heCeK2kzD18U0uzjAXjS6prL6J9VhDu0bpyj5ecJTHy
AFFJrfC5mf6zv+++4YpYnNSp2xIosPxa/bHaEk5S5JgZgBPp5f8KMtb8wAzs0tprKzQJ4plFy80Dft5ccM/Xv5OPy+elHi1c2F+Z
8iZH2YzamgP2cnSmnrxQAcCWBCgG1KnZ6E1NUlNdeQ6ftJNtfKnXnsfXEikdF5g05+em/jMuW3dlfRQBhJpXPn3t7a0TdjBYYZqu
+RJeuzt62PY7vEwUr54TaabgX4zlzwlO+EwVSJQd8Lh6XWhV9U/KlyP0baHltJ7a7vQPFUv/yJCJUly03AkusRjPPlr+LiDfqabS
ZRO8VnR6bSpkp1w+7bsf08KWFNcvLIAHGlwHaHlTcJZqi4Kh/WsaLXpZoIDQ2X5ETxDwU1njcTrvqO/JXAra7PIFMbZ29hmJR0C/
tS2uDl5DAot2AVnN1d7A+0ofTkoI3iQmwPUR4l9bDDJoK+3A9LJkcV2fEuynZSVNukF52S/wt0/UmK4OpZLKM5z+D0wfWCaO4DGg
97o9H3Z+Ssy0f1+7Xv5hpTe2Si7TVq34O6Z4jXF/+mvRe0nTCP7HoU1q2n+/pHDi534UDqs4DwDaodkZMtn8Nqe4nh4tBzD+72+S
j9R4eSC6PLOewEm8UylMBI34DWx92JZep2e/idEAIo8g0+iDgYJzgHkAsOEEgVZ3IZ+uVlqmNW/D70Deyvk+gKfBbX5eSt0jfZrL
kEaW4s+MbhWQEfnfUHvxMU/94o+aPLkyGYsTo1t1w+8vsku0I22PqNvQ7blKrxz2ssba4wB5tdI0A+kPjzP8VzF2gaX8YlwlawMV
yCfeW3OGGAY1fAcvLtwHXEEAzVp7UmkZe5xTTMFHplqdUnMEOhdrwHfJ2ypeZfn5Zi1HRuagv7MjJyc6lkEI2TWzrzemLUwyQOjx
2YciZ7NcI28nfTaJ91YaT5K4M6YCdHunQ8TuPkdV3xm5TAv+LJYPgvnxT2am9AeirdlR2hCI70atvZfFx1w6/xWXukNtx/SBKNKS
rTk19CERrBO/SQKNFmZA6nvpj83rcBxmbjmuIlJmM+AzivOXi3JkAfzlRtEiDb/UcTYk8CeXzaYUs3s475oyJsE8g2GTdI4/qEyu
Bz2d4Cj43crPdIeus/LSRk2lPP1GBznTQdEvPbXjAj1pBxOQVAyU+47X2e8Zh6h4RLM8uj1bY9nGfAnvMeQ2C2sSb9xwAK0ZpDZ+
Y/twe+6liWVXabpEnJ+j8qp/L4vi0SWoVIuAhyNF/tOKEHEum9XcrBOGMZO3PIygUlwCEwDRi31+w/Nuh2buKma3nxtnjck/TUNS
JxNtxA8BZjf5S/AC6XgE4MKMyyzBaROp00Z53PIuPMitVggDrIEP6T05ZEvqAXRyp13bnHO6GIjCLX613pZe3P3/yFi/SE8z4IkO
U8N6plMV6B3k1zcLqp7J4Kjb3YMF17EGiVtcjHNqt4IQhaGcomvLQ6HRfWHUaJf5v8VyY+59SsaXkeMThBpp8DUU6SjlGS1rtifu
rsJWG3L126F4XnE9JEF2rKUQhTCgKV39VnGWEKerSZx9o3SUX3fFvQyoWrhhbqZHdlWtTCcwgohvZDp8flwuTjoahhoi+HCg5YG+
dj0ka/umIeQoFGXnyH5Avkn5fkQ6U7LNdpmSEgSOyoVZ1n9ZR9KjM/QdrYp3fF9hAViADdTxg47F+bl7y/32ZPC4D9rOx4VOI5BG
KLPQ8uDEJbaqtJkvBGwibe6/VMoB6k9HovZnqitP7KWRipeVg65UucHoAFfI7D1l0kTplChGDl/EGPpUOTgTx+BNrebjvlWgQooE
j7jyk1VXho7RjExXKOnClc9hEkk4KrW5mBN40difZsR04l8Qxgdex7qSisCuHupJdBWeFfu7HXod0gxQ+p1hGS277HR2K/25o6D5
gAJWCfuCKBB2m8w/OhpouOFcDDnqhLjJp0bhVhv+EOAHhnXwYJ+/RWxH4GxnR9OpBeftbH0u5qXnEgv51TjD0af9vaG92YJ0+wfn
pLbVL+XQskZlAQERTPjawIzkNjrPQl9k6Hy8XuHl0lxImcHAMUU7sFsPgHGNXrqu/Ads4w5pYS0Swkd/j248X71vZc5jdpA7qcWA
GMJmZlUo2tkXu1uJKv0viEUGTfgRWafczbLv2mPawu+mifN4N+r7IsE6q1Bu6npkv/agg4AspEc6qRh65CFJ1otcqWQb57csYMax
9YTfcnUIcEAYVqTK1oIq2lPhedX/viCXVCLb5XNze/QLLG+Eu40b2YXRYCLNQxzyLXyMW/lS8ULpK6hOh0vsk1Q49UJLqY2+v15W
iXq3i69QOZLnzr14NggHtIpnRtPiMzPcKB5+eSKQBrgdw8vVDAb3WaBVw7FLx1JkRzN7LK6466UOniBrIkwYXIgWnKyY1IIk+iWN
OMXPrsh/h8ig34w+0Nrce1BUnyVvKVBYFFHSBdnrSBg8Q/txrz/Vs8nfc/hAww8OM6SVnRr+HEno1c40JshjC9OP405XfZQPbXP5
r15xUPXTFJ1XkjuL0NJTgaWYaGonc1ov9gkK5ADp5bSRSNSSvqtnjhls/8mDsfZIigkUaTaeSbXD3tzwVTKsRIavoRo6Idx6a+gN
l6rP0snGYy2VUnnqDWW6vawRDEH26H3SXw6LxIgMDrcjXoe2Bx8E+aovIAXb4lt2fUSZJE7KthnIrhFM5nsJWTYD19pXtRf0MtEF
WI9CPxEpCOFFnKCoxc2ldPNWRH5k5l6Hw/VbdageivOerz4LENNus+Jus8HWBjpGLa35d/E8yEv50H4kWLf2XVX0UjdTlct8hX3J
GdE7LejUzGBC3n/1p2yeqhozIi4dYH2EWJT1sOcZPzdL18RHQiBy1DS0KBcGZ72k7bUukZxYhq1OH+Gb/++47CjtGQuvsEzUSqEz
mt2lV6vk2QHvtL6wEv2iUVMevdzqUAlSlUJfnN5pudT/FrqwmfOQubmEtajl5VXSxjTNnH65Grj5ErBcT7pNE3DfluO+14Devtz8
UjQgh0y7bw38twIkwJf00NwgnzWH/zpBS+Gnvt17q4PC9TmnFqHNP1ihaUzsYuA9ThBPXS7IfY7C2c7Z0dJCEhHly/wJyu9NpDQF
nT5w8hgKL/PLOHupvIDCJKbwNfPL44j9PfyFqHKlQ+9LyJLj4EBQy+oZkPFPMe0CU3hS1yZKrT6IlT+kFHEOtazCOz00OlkJtXdh
GgxbdvKhhpuJB9YbH1uEGxGx6SX6Xd+GpxrnzRPMYCPsHJYNroJzTfOsnPYdtewOx9bnU3IeGOM82Kkkez7k+5t1i682WSCL554c
NW3EK5h//MlczFMe9Sg5a3QSYPML9swhsoaZILQ9TfZEiPaUxmgLvhFh6hogVRXPvFEpq6la4tdQ4EogZMpNF3D/8yrMbZVHbpo0
ee5o5c5IT/WFL39t44C0UeoH+PgtkpInkwWZ4qrzjPfve2OHdvcBl4nGrU1PTUCEHgK9sxty0xFP/MTsjgaQtf8Zm3/XpCTXZPqe
RmAFyx34blhVSY7p2TIuSku/ph5IMfOYfSHkxf9Fg2R04aZwI1cpZfdlU1amzh6QKaqr30HvEkFwNljf3tJBn3ejlCTy1GFjA2wY
+O2R9c2JTQH0LYzUDIeScjMi6KvidDkqL+r7/IIfiak9pM3BvosErOKYF2314O5rn/cUXj315q8OTf15IfDsdzA9GVbgu354n+h5
bUg53xNjJzNaldVe/2OKJ4tjq04opjsZjQUa/tdSElSo2DJdRM8esvHoRoGiowRPJNQaqf4TSDF8aWxR257+0v0jWAvM4BmLAs1F
M/TLAbkotuIfCvX6j/bGqNn88lwKcMJ8cUrM9ddBzpvWgiCKdFrPR2vkwUbC4qm0aMEH7WO++eCWqnpsPfrocf6FWVsTm7hJjHH/
f35uxprVFBe2VQ2aOmkBY0/ynhCRE1h5iBQ7kgYpxu9L87S8qU7+NdZpncY3AF4q7FoDWASSf91IyrjLuxVbMqa9sHMqbJUx+xcg
t6Xkd/x8Ys81EyI35L4j1toF8QNFQYDd47hlVnK05LcJtiu/bxk5vdtoeubmWEwWykZaGaCHQsrElKQDpdA3gtGo6KPDDMPgiUni
gpTFY89PTdD5332WgPvWOdue3SUBHy/dpVOgRd/LO9NVzF0SCxmu/29AzYzgl0C1afneFhmnZuhSazutlG/k1aFIOb5giZllAbOK
21akrT0iYBvym9yyZ7v96Di39bRpoXuIyH+AgD9b64SZ02iB2YZNdM/Jv2TGd980yGvENj2nEhJm7tH256KaCeBk/CYo6GhdLF/x
T8A2h60+tsm2jg391KhYjasxM9GqPYN8Pf7OuL4lCEctT2V9Nl5AL297RZBcRygxlpfJ6m6XHg/LUcNFyxrYfzmFHddFOq95JXhl
6mLw0iMXWo8l1jVH8gZEFBs7drHT74lSZFiieKBoMK1/J+0Nt5thEGtYWGhqG80uGHpu5gL7i8y5Qy+++VPmayWkdWP8wUMNXohK
6L75C0OaQ7N1Ks/AGOaJjmWx6Cm7uTbeNxJTF+GU4Kky+pSV4e+L3uNm1Caerrd7oK8qtQ3llAekOY4L4a2EBWqOT+W7cpnM2Fup
to9SVR2Ta4Dhhih2nBIStzwDXLRbY4+l14TJP9MvEnU6Bmbyc614RDEOB/7+Y24p+HwDFzP6/3yhzNHZTd4c2y14vKimwBiUJZ4E
9W18Y8JoEl4gDW0n94o1xBnieacL0Xy8Qt78ZO0GMXkX3CnZ2mRi2oS/1omi2E0WyQO2lftzrUFYzn+VBECuqKr9PcwBF5wTlEz3
LmVcZbrAqBiFnU0I+WqxVTroji6VZ910wcZxicy/6nZNv8yURui+BmBqA6xGPwLC/BZyswpTfhT5YSs60x73oz+tyxeNqVGwrj7B
hJhiFYocM/ryrzGsM4w5P6n6JdIdqyEDLLLUslrkzmkoen27nD6Tgkk2SuOf4LLhglLbruNyp72NkUpZk+VO7cz6smSCIZq28/Vg
ShZhXpfP2G0JaL3PHsWpp6jytp8OT2ZW9hyxdQ15Sx4dUSUCrQAA5BoBhCgiraRqkTQkKq8DsaLvUHcF+L+Uqyp0CAvLnHF6Sisk
mUOryQBLXOMJiihmELZ0qRAWJ5SiG2hsSyNuTXZR2etp//rbDYZQd6ZRrbdQNJdmQKxeNcdQmkVi4hz7sFE5X7nw3oJI42F7sht+
GipT6s1Kmvw+or/DxDZuHIAWm4B0KENcgi55EoK3js5smd6sPUE07lIxiDP0PLEZzT4+44WPw+YqQxH18H//VDfeNlAcFNDwfo0l
cx27VZJoKZJ1n9qfwUdhKoF8d+GtBIXykvJfE6rIjPXZh950f2qxKYA3ZRbPIOWD7lZHQ5CExPzW+IM2B0zQ3GD25wUg9jmQ/0Y7
f/rnXVjFU0AbfRs+5LgigfgHva1B8mTne0cQyiRtTsRa8lh6bkm5rPHZZKnL+aMLcMT99qTqcD//5c2dq/Qid+hSwz3cZxqtLiDJ
IeGLmVFGDemABwZMNkIa3zGqq3agsRwZ0CslUQDgZb1AAJXsibdOppZ11jjXSw+44tD+HwNCLCgTB4kC/WcuSRkVuhKs5RKf5Ojj
7hHS4dkTDOPlpUoamosE/NmdlODVJAKUWXS4Qr9JDSdQp2HogkndElg7XQkCtQFyLoT/bej7712WBMmnFIl976JTsipsMF5BMxez
VQLczCGyQPE9+5C3pjFUE+1C8ZwjE4qTP6I6Ohi184+oeSdRni0MdBTlUT4wY3SAmwAbqL1KjNrMb/mm/ysU8SVJypEejGPzy79l
6uPzM7G62ngEoNIQRgmxfAk3oZ5/pUqac/u1YK7fGp5VZLOlZp9C9T15PsDP7Hab7EoThDHXYWnEggRlAzyIIt3cjD1GYfWJ3tbX
hveha+FWhgWvb5v4/Zk36R30dc71+szPAN88o5HDR1M6zs2eGluytRGeJ/GJfwBbc5qA0ELzW0iK2JhNxvMWOtott8gP6lPFpX0H
zSV22Wmb1OBGoDQX69jfGlt3KB1QRYpGaR32EpEUOe4TR2l9Y0vTuUDm9/xBvVKYfyt9evjF+njEL2JqE9TLyYYyKf6tHbjJrgE7
gJofP/n56wGTjqap2cbGCRxrBPi05yWX2ALg3xQPMqZalkvcW17lyZ3jcOxOdqa7uyVHTNvMeMwisdGuw2bkeC30l3z2RSqzQSD2
bsErjUOdYgioDhhr5VyX/9hNmgbG30+cAJfM5RNlzZgNKwQfiDFqmhV3DHogQDZlAl6rdhg+SRMyOkvCyYJ5+6B1wBS1QHAoaqVY
KceshIvKDfgY5o8N+o4x/Dwep45pzzjumUATc3zqp8HitifV1IijX1rbIOG+sXM6RIfW1m7FigbSdrr/CT78Yp7Txq44DWPuyxPn
k2w8XNNQXtFdfia+8tx9vKOROHUWVL71TSczQ9SCHQPWPdSBB86LtrDx6eT4Y9HB9rFblGlb0U3AXAQpvuhuwV0Zcjipflw+2tGW
8N6SgWIj/evxV24W7AlDhkU92wmPvVwbTQAyBQ1bE7kF4QI6xXnEdtarlBZGMBcamPKPLTkR31zgRjjA/V3lumv7BcRMWW4OkMaA
9WeJcAuTRjU1Ckb0a/A0ytt/1UNEgG8wX2fM9OqEOLLPD8eRdqiHQzYKqCbh4v3qYbzU/o+uqkJBbmpLzIGXwzWiHDvy9SrtavEn
cUX+pXkD2p5kqzz/vwLnVhDFKwkcU66ftnv/l310uVglIe6iNld28aCMQiHY5hmgV7glGbh0Wqo8uw/N2vUz5y4CVbhIAs1R38A3
QV8idyqYagXsDVEfJn80RGtlPsayxdV5jAkFlNUXuLvvdimzLcUBHFZSKT6VIagOKaIDR//xaBWKZUjm6U+HRq0mQUa/sc/oSTqb
mvKt95vyRZRE16YGPUmrpiLJw/wBflY2u8pmkIzG9uCx6rdSQpDuAAbC1WT3/ls2+D552F41lrqctOr/3pnihFiWl98bE+0Q0399
GRCHg9colodmliZKHFr/+VzO0rxLNxVUb2X9NuVV6jFz552a0Ev+1WeM1ztklF8wCCvY+yBcVIHcsHTjVfqybbPkFznfv4kw3zzM
Xywkl2wAJe96Rb2N2sUDwPKMReoEFX7nBTOw1wHz74RDUSJTNYjzb63hU6KJ2CAPspiMuThsPKXvPGKmZdUgPGw0302HpQ/2t5XD
LwOPYPKMDTT7kvp7EUZSYmmhwEs/X6WdtVRGeS5YrkfuRHufg1zjCs7DIIBZwVGjVE0bBrOAk+EDd5h1tCUxr2wiYEXHidicWVha
Vwd1a8odDkCCSh3mBEZhW05OSOSbqKm75aXqK+SBXIUGTthx0dR18yHb6knoe1zOxGxlpn7Wu26Dy1hCjE++PpEdpM8VQlSNd+68
Ad81EWrGkbrO3c7ZHUxRjZXsobml34XCZ/vP/4Xt4NTzLPYb6PQyFobFKiNnpzZ8TnKhctEu9zOhPHskGwGM1GTt+U4f5ZptFah0
xHnBWPFBP6QneYollC+w+sYMI+nZiw5eOhYavVP0eHLEOYirkZPMkwB2LUL3aVPWQfc/VfZuQOzUqO35Ylc3dSjVkcE11GZUFT5C
o3BtjjRdQxLd+9r7CehO31e9fJ9D4m7YeTLVoRgbwcsSN0UBERJYy18+SPCkRmnPBcLx+UQjDVaVunCvBaKDPtj77KVVMSAoHgc/
B80fF6finv1WsXZK0St10DN71lqz/6qTMOi9DEc5YJpgkmYWVnq9SkhLlStjKI5//tdN6SLV/Y5DupnkVdrqP8LNluK48QGh0lsT
U3zQd6QfZ4EZ/ZCb+/mRLamJgacrvaEirSMnHlhE568t+Cda9zF1ccoC74sNZ7BiUkM7x0r0gws8qKg0ZJGjAdFqtFxzW0hqe5oG
+XV4QF7ikNMyu2krysZyGChrUhXgWMwHc5kKS4X3qm1muJJhQu8w0S61MCd/dSK9ZR397MhGsHcVlG5JnTKkOLGzbP2X/INff66d
t3vHi4GLqtc5JJRIswXwNR//1MA7HiP37x4a4fd/eZXQDuqOeRar7YEOpE1sK76wXJccUVjKdVR2ltnn7y3UcV5OV3Ivo4pxg/WQ
gL5zxTb+CNeHC4Wgw6AXDAGQfX4NpbA8ZuQzIL0JkjHs6yrsxHPtBP60nQlTnu7mSxH2FZHa4ULA930Cc/eEFNwEkDnfqFBAzpG4
Qx71XV88YX+WnpWurwrvdjMzQpSN8VxMRzFPVm5NNR/vOnUrOz2U7m7U9QTZ8t/AgG6z1KJ+7PV0WA7ECT7H7LuA6khFobuYBBqc
61cKasJFG5qhXdOPEJ/5cI7bRY/kuuee4WY91pYp0UIxOXcjZrAkLDuc6K3XYJj+hanezre/BMhBJ8WIvDCQjE1OJlmhqW6TqFYy
AWWkNQdXAwDZuruMDiBe0c9IhzyMnmOrcS4ndadfdi6mjKR+xxNDMvhPHoeFu1v1MIfKbXZPgZ/FNk3L9lxwQYhcwJUyCzcFN1VK
5bMo7ydIg+g6FYPNco3wXvfZTh/MsBFjUe+RWHkPmAY9EKUGSlDI3nzc0K99IJKP3qG0gbiNRw4W5JUG15vK3HdKvdjjdyLgGesU
tpVdNiwdlW2rWiCYkhD9zEc3HiLIQUcR6oZOqqo/kmqsuGw8uaxbD7PZbHzHhyUTOxhq2tsYdVHsE/vldRDyNiOpjioRkOXFtwpZ
xphqz2QCFo0EXqb2WMnEI1rpuWq+3APVAsuYveiuwn67V7oP50nheGr0cvg0JjH++44GNUd2wO2mW7UTrshIWvW2JTPvsISC3n8O
newfJS3rIc5cQ0v0lbP1yew1QyxitCl67Its2pwoIqj8ScfFX36G8tvgO3DCBmzHAj7wTSJ9SIf0kA7fw4oXGQZtXrTVOV1NQEdG
cQmkHAMLE0h3/84yAIATKDbBJ468/zCECnHiDAOlMMBRl9a/98aHBYxJbJxQi3zaSNh/SCZiL5+tw5EE8XVu3pmVR062HnHF1q0d
0hGMI9Z5m79zvL59Q92igUPltHFmuKRQMg8pBKVkMVCVrBjZXbhvEzUVd2DpIyjwEG1tChkCidCqsViENzVH5SdNIMlOPHofqESR
Y93YyFY4rCExDgPyFEx29M92poULdkVaDdChZovW0U0m0IYhcwgIrI0zMwj/x8lKoAUkVc9ZMBCb3acOMR1cYFef1iz6N0j1iPRD
M7lEQblJ4lYAw0ZuHQVWquZ1BlWINL3BozKL5XnTe3aTFpdG3aK/5AeDok2UgU64skrPNc+ZZHlm7nNKYj0m3kweIC6M5XZibgUz
dzWussqV7WH2VNVGS0pdQnNHtec7mwuRYSza1tFMLdzDLijmiYKh+tL5O0a9Jy0GK/wSLu8DCFFvQOTCPJxSh/4LzgF3HJ4YPaI0
hiGF/KOZdiZSL7Sv3eTog0YvLDIroOtPvv7wlHX/U8Znlu2z/1Uw/6PbQWLfG3JgIorWfwH/pSv7oQ/AvUDfevMqitBHYtLPuMHq
BaoFZkfDgYIL5g+O1ra8AUpq3NJI/So4NBKi4JruNt8UnuljCXkcUmqPy/eq+EKQ7M8H5wgVHVucIjxPLH89fSQo3qDglUk/25VN
hDvygKdMX7ccuXOQcKtg46yP+9JjYnvCAYEPbGbWe4PJO+NjwrQcRnYbz9r4gUs3pweW2GyLw1RHPj4lCw2HZB/K1CBU6hShS++t
6oXb0NfFmi+1oeeld89SY0EtSqAtoLvoGyVj3gL/x9tKTyoUF4xtXHrL1D4+GLR3/zbOfV0LQcW9MjHanrHK0U3fLtx3RpuurBuC
KWnQtWe0GoLGIa+Zjtibc1bZ90nspIvfIhoTKYRr/DQtaE19ju3E1PS9Z/4Hok4J5Q94t+8udLKtGaOnFBoYpAki/3wT2gxUQOav
1V175dV4YOGVVTjCVV/dyKMnAyF4nb7oFTcGik7L3J9ZVmWbhSy/i+/wiaKHjRXc3IhkmkTUHdao+pWsSiM8TRJfY86dgVQFXmda
JRkZ30kEVSmloAUYVHwAplbtkzCFGumG/BAY5M2jocks1L00fzBr2kunUJTgKYnbb/oEgSxGM0Pj/A8RLtSISUTLQpvaow6My9Oj
hdTxU98zyrNQEJO+J5FjCJ5iMxvEDFqfvyI4soIldqKEkmVQxb5GpFj+U8V1fvTYs45UCtDnANv4rn02eH0II5GPnDCDqrSQMEVU
cS0ORkmmoiUBTbsCUrcFzfrhzakqlaY1+ZJfHTlRteiwgjA3uQ/bxTSp7yLVcsFztKsk8FsUwMS05Wj1NhNIkF6/uOnLOcGlkWyS
7yAv+HbXsSl3Mh9HGfEWwfeyVWVrSbxULpwyqneRupPdF5HIwTKr5y51vL3n9a8yJf2l1hUUViD/QaKMGlc9vzG1p5cJmJJjg4ov
pRcXRev+o8HFrvcM6ryNDvck8x5CatLgbMJGhLNydS/qEAsRAp9fzVBjxpFXkYyZ0pilTkSjtgoQJMlvdrjYqDNIck/Y0RvFyunN
gqB3Yzo/+TG0/9cTKqzrBCoUDzKss6dUfgTBXOE8CNxz8ni63tLOhZMNoh1Cy44oGDRXFjTGjO0jqNc5fuWYLIHBI2xsoBEBfyLe
eIIRoR1p+GC2nE+l+U8qs6A4wYAHSn4fQf4wR2hQ8rorpU8+gYzhMZNJnajWwdIqXfdhq4yjzCguGO9fiNZGXdHq+6gIfat9wOsO
NAhXOTIw6FU91htMsUk6afetrmF0sjpKRl554vrLv1Fxu0ucr7j5FdD+Msb8GC4q6oQdZ5l9gJZ9Tw6fLwx/d983MpoYemXqe9YB
behhW0civkk9WMu2RzGqsdzMRoxxtLoBOpSQjY0MVpdszv+Fe8c8eVJzjJJfMvbchF5AHT3g+ROH2AbppXXJCeo/pXBltDKUdOlI
Xz2qeaex43EJxCErZ7o18FugyfVrUikZ86gu1sEITQOARUCVJw6UPQezIuGQ/Ttn88e2MmX62hUcFmYEGQFtolTtSKSGzzxNSLHV
tQ4CsPOBiKUDp4ZsJZ6o5yYEL6MBxCPRpTTEQhamDNRbpzMPCT5KcJPKo9SvtGNd1fcFztEJCJXDMP5UWmqpKOUhejUcr14Y6O4q
bRykGE6ocQcFzr4fGatHNFZJNYSiaFP02eY5WfZC/hH6+5tho4FzPWDTm5OfdBsWbLaxhQFr24cU/1xTJJl2yHr4Itbbn4X34GJG
fhRx0ZWUtmtLJuCMeSSPyC2YuIJaF1ogJ7YYasERpY36sPss2IpXQH+PSB7wB/tiiy0zKNU88Xh3E4WC7OMrxs4aIcImFMEwaImS
Sl1weTJvCzyfJH5/1embABfyGgyMbhO5QYqib+aQr5k++RFfIKqisl7Vv2tM4j6iNHbD0iw1SXI4JCbsYyKnd5s32TElyu3lKnAD
9DQsf+MMTupWojpz6TYNSG7cwfPkhjhfb8mVz+Tj8pLCMC3M17F+ljKIACFZxFxvA4DuS7yU+H2O/lA0W2pGKrHPZt9ZYp0XCeZe
D6RlE99yAqCxNM1sK/wzeYbyMDr/TCQwGoMuKyYydw+XyyqsaqKLl+U5X0+PlJjfsoHkQ9qAN9NK+LSPAzDqxPUJv90GuIS7BpZ+
MgjnVTKj3m/yO85faow14BirZ5DzwHcJ1PdhOPmrXgSwkyaJ0YeRNe+k0AAsrwRuF56k6d/LjggqcxC6lyoaP++ebqV5PQ4fyYoT
ATGS0bWRMLqdTfqkkXwnLrHz8VclXD6yUws4L/eHaJvucjeB0y6A6Kf2bcXOPipcYdoKQj0luuJ3Edy6NgyhkG9WeW3zr13ttuoB
u3XnslMMdtS6L6uJd8lUI9A40ORQKvcte08rZfCxRWmPBE8iqtkfgfUt1N/3N4nE8rmySQIOhwNUMo9h4Vl2u53PcqSFnIIf0H84
tAOZUGC5qQ3gE6lAGCCkRUFlmIGU3fjPJGeTYTPrjXHAjaqNi8YupC16Rt3qOxJpe8zg6KBCDm6R8XgR8CiqQgQtMhExmMAA4ADq
TX/XMdRRqf51/Dof9zOkQzH0r89eq1oJ/YRHDg2CB/RMppSnpPcdXVoKib0r9IuI7IdvlH+Huxwi5T+NZEo7kl78YrZ+9KScmZRX
e21zsnUbeyMGtY1/HdBSg8tgAAi+LSFqR3sAYZIGnJYOTZwVB8yLZbyeICI0u02sBqIMcZfuWuRGVHc/pFo4x0f7/tuUCBVuEi11
BexwoPh3ncAjdVJ8lUgvX10cN7VP5RAh29K8pOcizUU8iChVETWnEPNkJTQiJU2iCJ5Vj7oaReCc+W8EIkPjHWxdmosWqTLAtVUg
5wR/4zDDayJaW6JFUYRYggp1qB3YK9x54J7cW0ucW44yt63FqvoO3P8g7CAxqGWYZYVrnyZC3J60l0KnQXGH6UMb86GufoekK9Bh
Wc0uAZq+EEazHhldQstAdXX0fwcKNdlFal/ZkYyCsg4PsMoODtxEu78usrgYp48yz6vDb+m0xhOYFK+0kXqjlaBzn0WEgO4K2EQ0
b3y6O7XoDUIcEB0LAQc5SLxDNDiN9JEeIK9r55N0c5vxJejGnk/hd8W07rMUWEpqgjWShGL5yyWlUiv0VjHSIx9a9YfAvnyr4Zhv
WkuTl5APkbLKwpkaX77EXGTgQRmjmOLbyQAmHQq78vnHvv2BjGPcTzoxdK/9DlBJ7wKpJzHiWLBOu3DfiovBblphqNEDJykCDudy
TOLVNsxLECIqxEiw32z9MKSPv4kHq6BK/gZ41wafFW5Woa1eEstldpJaJ5k2OElWZLMTd0ry0rtawpeQdjzXoNy+5w5j4v6IeicG
fYg3c5GzIwTaR9ZGSY2xt692V4JXrJyAFL7lThMJVuVGwiHEeRmDjXiUDKie8J0dSB4l6NWdAavk78kNovjUfa4+YLnMv6hvUQ/8
dI1SzGXuqB5ZGyvYuwo/VFSkStwMVhpWx/15QhcTcTsEjJN9QeYZDPDzTWJweFdEhs2FK2Ry1TkoGuEhM5bmTLbvZs4an5jUtWst
Fy0bRWbBY3dNVFcUoz880x/6lZA5cJrJlnWJCQCQ4L8u9GpsHborY2/1g6AEu+0C5+8Qfe73NxbJMR7ctO+u7hPPa2+btNM7Aon9
JAnxBYbPGW5O4WslOVwByMChn2zmrArQTuHZ6D+PQQjUsj2P1nTXVv/CuDg/vP3+Lu9OKGDksJkDU78BMGHwSAvtFYArIXQakxje
7I05LOnXzKZ40hfElibZv2cUqlQcJohMc98NZsrCz9ltUWSqOsU6PfOBriK/hkBn35xx+Eq3XeZ/uTKAGRKEL1B1YDQH6Rax+mTn
x0sl5H+CPGpgxQHbrczld7SeAiJEiknmZUetLbMPFRwjdExC5C8pv8tkmf9fXwpzr4N3uPq1be2C3J4Mas7E/WRkWVScFiX0b62q
LuZ6i3S6iA5GHQbaJh/IvVQpDnjTn27kmExtE1N97fMhIXAKivIlcJ9YBh+IhUYVpKcgGPxY2g+FN0dkny8P6XnTQzZA7jPJnjkQ
dFHvq41DToX0BFqldoSINS8hIlQXDQjebLtnI+OCpM0pwefguDtgppTVrpUZ8kyvVMKQXKdXk1gV3zFHO5rotvm295jLTP42OXNe
+upuj3lKgYLQfsgojoqvIBw3P1LvmxVGDf3sxsBRXmfVoBANGWB5snF/27xeHNLAHkNzSkN4bSktpO4/o/sKpNdrbRN2jlDs6zEx
SVDSYD9yEWUKJc5kMXth2M80hfWjyNu175QmlHnpKUXxO/X5RRNgxyuj/zafxlCN+Wn2Kn+kBYZS0qK/0ne7JCAKLFIPkGlCzuJ1
uHWXxopePAzcpeHdzpGFOWdUks9tr+JzQlGsn96oba0qzYLM+c2OgmQJnLdzm/1kuiQ5WorGmv1DKlpqrCCo0J89YRVDV2West3/
sd8HNuMptf86MFruyYnIXmjTMc5VlmuWTZJ7y/mR6ScFk+CW6RQPtN1Tk/+ESWtQRmC4QuFcyklf7UNUEAykwiwNAS5mp28qdUB7
Gy7Hy5qWw9il5kUqeSKDbJfzi8bYPNdBChOtud74T7FhIFc7G7SMOfiC981OuUjZAr8IZ9iXF4uVSH7MV/tqVP2O+V1v7ymJPy/A
6oaazYp0FKsMwtXNMqdmfemDD6y00w913mxzILcUQEHXKqGS/Udu4HHjcb7ameTQNQJNoGSqp24+9o0UlOtYHuKgJKfidQZ9kMqu
yfS6BLfRvdNDXtjoUcJMqVetj+r8SVkQZoaYDQm6MSW5RNH4B1wf1ipafP/sfaHbsltHoB5LNkIFePWgaAhjiP7Pl8j/TSwmjGmU
zbiyEpCcCFnp7yWj/G4nTdlA8C7LNgr4pUNti+rj14IlHoVCT1nJP5qi1+uxf2pAN7r+L5SKMg83iszHSUpXgSWx/0JFaqkMBdN1
CUpO1vZFcCQqX2nR8kb0stWJ9Hir5wMf/pGeIfnFZfg7ef0cvbBeai+4cIZxN2lGcQljzN16Tgp/Tj1DDZM3PSwrrNz3TNw+D3I/
nROmobM0/o3f77DduQ9ZMUOYg7RG0K8Totgg6KyIST5pjeTmDe4NnkCxEqD7Hwbeic5N8kCGDtSs3yXQVfeSwizZFzHWJIVjNdqc
vkHvtbid+lAIUVf0nEZpktYGM+oMtQT2rNwscQ+nLuN4ESNyvM1BhVen+S5xw1gF8PZG9TXt+30QMmQ1bznjCZdZYFT2TI8154QA
a/eKf1aLfhfONMJXCGh6SZqJ8gSh+cR3KBMOWK0k9ElOFrJ+JyYj9GasnTFmdiYhU5rtZ/WmyMSEVrb36uCCS/s/VxGLNRkwnYlY
N2GMoRxeTuGb+J7tt7o+pe6a0pe/WAjkGYN7VV+X5P+h2260I+hRiOy4Mqo/3kVUB4BulEXvGT3frSKGxqXuY94hri9DL4BNv1/9
4DNz8inzfGC7vWiQ0gQ7OpVb6AaB+GvMUsbg07Y/UTT5CZrEHr5Nq2kE2CBUtg4KJdgbT1XZ/3DT+cc1ShACErG3PvAFJ/iYP9+g
SaNOR5yfniBVV2QAf/0VcpOf5T8leeiHnhEGF7n/csgAiTzBKy2S6bAboiEWZAxsaalXcLMUws5+PyJOLDDFdN/dVwJ7q301r8U0
cu6fyb7BX06DyYSw4poSUdBs8eNU8mLf8MEJR+xmpg0l3U+MsvK6jE29/zZRSpKK9ZtA7BGtgy8YCP3ZSELZ9GlW7B5B4FL/c6qS
upWnRCoy7O5HXIlnlGlxfFbz26ISmX8d0p8c5rZ5H8yiEpSJunxeuZnLMndhPNSuCy+RDMBlHevTtMbEiRDHyOW/q6XRHE7LQx1V
UMYdk1tDxFbiotK7jfPZnxJSxshWhXfIU714zgyYkMMp1ccrlnrgjlvqtMBkVVxs75pUXd7yI7OXnni95pAhdKJ34av1H6rFyjrz
8DQcgt4dwQJbCN4fdLVcg2cr26BzFd7vxEb9qsy/BOAQh8so1L0uwODcDlcuCYD2kRQXBOMVtLu1PlU5WSA4G8bXe3kmlUNBVLez
7ALoE58O9lUIO2awvyjy3OkC1o6THRQgj8h8iQrq+gE8O6OQZ6lR20dKXtPRqj16La01OswEY2WclbUYh1C376GFeS9uQ4utOOdl
wK5kmNYFQOFM2sK53s0v1oC776rbi2i/VLtLrksssuOoDZAbNvbZ9zQV4U1zja0oTi33JdwCvdwhhBaAsxrqQpbg8vjwaGqp0z+R
01agu2vG3oVG+ioTjYok1gxgIt49gFzKftGuok+9u0HmR6mLg8Q2QMoM48gMszxXc8sRBP1+bpU/ub6R4oCK0OpkMufREmG0nG08
L4maF4K57SseVMWusJN9VG8tMXlE9szmirRdWw7W7rc1rCWHQHTtQSU43D7r5Q1UDnNZKKT2ELG77QTsJCBDoCOTRB5L3aa4RRY3
K11lhmFWkdJdrsDtTpd/g2TUDeCopIIL/9S6LUPOYbp93YOEV5V+a92fOldtueFByFr/NVNsSQpjNX2lWGfur2/Knp/ze9Lp7cix
BMYbXbeyzTkN9AKF7uw/fFdwCRHY163f6kJxGlDzpz4TPQ6OcHccgCpqHmWBXV08g3NFG6ApAKUktYtkpfaDzJpC8Vprr42cAjEH
8h2cPg1PY44clgCVWyb0BYAcKEo6XJ9W7hTfzCWn0QFYDj3j8HEAmGd6s4Ew3S2E16W+VApX9WMWks9cAoO6kZOkmchc21rXd30e
Ay0UU2udkMJuo64wihJNdog/XgBq3gJUST6B7YITtUYKoYtibJ3f6/lsB0HF2xinSyATJhnxKCdCwdbCRCYE6Z+HUhyd4ndiHCkX
2eQTb30DqpwCgDWW8vvTJ9rUdxyoI9uZYOjTrFrwJx8nlOSZvujFd6Xzrel/xl3stGwUQcW162p3iw7GPERTWiHG6aaiZZ+1axIz
1yDByCBJyy7qaQIVfoawuEgsCa0LwwrSHvpQayf3h3vJSmNN5oPnKHY4/1Fo6XOH/D5Lqw5kd6HPh4ymKESlsRFm82TRBIdF5B3N
p6IXUAef5cIAIZrfmEvxb7M0nQPBsHDwzLXfmgGbQ4k9qJAbdgOrjRVbfxzeeth22rSax0nJg92sFXwRSt6fKKTkniTzv3cN14j+
eSwuvGy2rd5CgfdepYY/lD+mVJYQlKEuOMjW0UymGc4e4o8iz6Iyx0I4MZb3S/peWNoXyq+1k1UU0KKOwrlDgswZQ1ByskOc9INl
3kYSPi9Cxd0pBshcHdiUiaoiFX/DSjY76B2YaFnC0omI4fd7tJNLxsuFLgC8fF/wTHBs9roKx6MHAUZipI6IYpFSX2MDdsqYjaKm
TbXsZncmK0RGPSu2DVCRgjF/MH10Y3Bhxz47YAfHskVDWRCyf23T8QIEHDXzwESYTTgYF1FEQoAoVsH5LfiHllaDX5nH8OPq6Jrl
6y8Qkk7mpg0jYDXUEAtsAHAML8POH0aYRfh3f1Q4XVh+4EL65+y5c0qscHrVEajkrdeyZXatBmWYgKL6RnX9gsGgtyewYsd83P1+
EJR8okN+LzwIFWplK1eTY9rudkwJELPi8jcC4QrJKajErIBW0NOqrV73yuEMr1q1/OuZk26VxDuZ4eAmwc3p/rBUXfYGNSqeqRE6
oyNcKljolamOw+1XQmNZpAQVp8QsY6NZI8yjUG0hs9xRP8DNZHRrptjHcJ/3qNxTrePRUePSCC4+rua3hlnGQ6Fy0T6fwa4rtn+m
Y16/4trD7SGqUzmhakZSdpnBUQnll46ICYAUKty2Q+gBpeesLPFbaCjOAzzvhqWsEd6niir7nSgurp42LVthvxH1ROF2P8crI9An
9D9fUm+ibupoUn3Mubk81Xb0f1afw/X1fZkE98+CbrH1VjJwpWnRSoxl61tzzFV0u3Ll/w2S/5CgLUGj8xpUdo7xlNEsWeWo/ycQ
2DJsz3wN+9C6ZP8OlQJZjZktQRiDXngK1U2h+Oucdpz6myG03nOZtBdVrKNz8/kkFGuFcB2rliHN9OsH+jvdm8FGPWbTxuhhyBnh
8GTHX76Aesg9/N/+S1qB7GFgQqrj560BhkYq/HybuwmFIjL/m+RfJRhOvPsczkcjyWTDSJudeodDjFraINjoQ8iSS0jYjlsXZIBP
Rhm+jmmqIS1k/8LiwdVZc0u9CxTJY3jpX86CkzX42uEzfZQeh+q2GzZYHiQGZCwQxdu8KkKJSG6XCr4WmYwbj8lBzPFFtThg7UQa
2pJDD6gjKPGNdOZBCDWXmPyMAf4Z+i2Cj1UGOiOtPsMoDOKqjPSzd7oz+Vqm43cDE/HUM5DYfjMxGtxXv8Vci1+lI5sqUSTmCL6T
nvGEKNWqwoen5xahNDkV4mOdDipxA/qSHWUQCuvKjW3Cvdyytyl4HQjkfhYkwhgKPIh5Fclg1zJwWKoFBLrzjLq/yqauVa8HaRbm
WAlA5G6aIvOBWAjKoKFBVZiPB8WQG7DpWxnwhuH6J8aveUndJzu4ECauyQp+Hm7SCzAcOce+7neZbsEEKF7vVZNrziXR3paC3qLJ
5jImYBHDLOsjsZwCmz8n24adS+NLsmJhm00HQ4bHN4bVC9gV3NXI6BxA67vMoxBd5N8J90baxRElXt8iOKyP7zjS8ZYAY+oeArhP
DHNMYEqAvO69pbutp7glUjdLy3go/60v1QYceRjhygQ2iBo+Paw7hHw2hEvDY2KJhl0Qp8QHD966IbN98o3JzVzy6CbTKkHQIHrv
6xo2NgJfsTijaM6Tmeuxs/xztjJFnJC7Yu7yuEJ5Y/EDf6YKqBpUsEJftzL0zX7JcX6H69nnG9w6zaLxmFHQibaJfdC4RF3W+dqb
eJirK/tQH06OVm6csgFrGEYQqfNjgJDKXQrE5BUWEnxY4Y5X0Gq4rnjxC//Mio/eIIv9DeGmkUbBnyXkSbdta35fBdbe4rYBmrZm
0/orqrv8WpAIfzR4YlrGit+DGmYbSWp2r3+nbMflBsN3fGc+OPSziMFx/wy9DXCRFUtQacbMKt653wadTHrjEAxAzBc7FhpP1Mgd
VyX/lcn4KBnWRiqKNMTEU6f5IWmXhcPmSM8ClE1GjyJEeA5odiZJ5RY82Za+9f032QEIzMeLqmC3+MhadDaG+zscEhociAQ6Fc+J
wAgrXwRg6s4mzk7vVzjofN5qr05UKq350V2nS3K7JAmt/p5h+0vKa0JZoq0Mhm9a+A0Db8ESzMfosSH//5cWb+MC9O4XanU57hRG
ydmiunVNUNT4m1z8UzHyMVkhznIyeeBNBzmSu0gNMrh6238tz/z0ZWoAQD6BafXBXa6obl73x9ik1seGnD5bP9s3DyRnI74+wLv1
Crdk99wpHsk/warsveYf+lw2kxrPEPLtH5Z1fNC7KAShVMEO8kpDp2J/btBlytqAkSYIPAqojUb1C8axH4NtrniX+Ornm9Rc36bj
nzbVs7lBjCKaw2SCekytoHMROZVPXFHGj9lv1HmSXojxnXiBeUFTwJq/arVrn0WhIHt68E6qL4edeSSXmZw11rChYw8wiu4ZqNZ3
OMyKh98+rRRYcX5kBbKnqgkbAK2Hhygk/TJVxujzXVZ642oY68WEUm6p4Fmsm7sQ6HHc+vXjcicPof5atyo7hXRSkl3DpXFgQqzW
SNRbFibre9w4VSMOw1MX3GxmmQ/3M3wtysHsaCO3+wRK7htTYKcsRmYI5UhNlITnAPs3QwPS913kXcdB38kFIcIEjkhlltsjx6Bg
B5cVkg+NnhOF26iYw45zhrsEnXvAf4AAgntueC5YWPB5m3eaFm0Gp0CbxmZANK6xChnIpqosiLjqXD/Z1Y8ZChvsEQBKJ+Fb+z4m
jhLAPn5/xxGUlkySESbvatU0e0W7dj9B6446Web4RtWqy7QPppBt9qIL1u5qld9t65UbYION7jEmTbVjuf15vEgmb4Ae2nZpeuKc
zI4F3uK925ulu4VxAtaqQZd2ahv/26164lorAkPLKq6OkR7pW9C7XkTNtnRRxq6bhx2l141gyFrcleCGqNzPUduyu1jV8rNecMxa
BdLU5VX21BR7Ff9wvl3fq34KJeLfp/GZR3D1lDOt15EH/zbcjSwIQMuFl1k8qw3LQpPcOVAPbpde5g52mriw3YUoxos/7QRunEzX
KZklUKYqcA9FlxQbESoRUpr9CjL9Q5bndUIyFml0urXAP3f2UkJ/eucyn+eirVO5t6GAnBlUTYXLoV50+r1PLCv493rfAfBMnkTj
bvkcOryxBeeHlINyDmf3aB1XvkunGG8GSmD1GdgM/ZlHEzO+VVLrhTR4z4bbyRNHyMCdqK35npplsqAmGrZLnszNxOnqWToNukWe
/i4nsbfwsCA7AUqSUG2hkORQHCStzo5+14S31J1kp6xdI7M97pCSwqlSuWQBbZImHz7MTz18uGcFC/34+x1H/hVw7/zgBs8vdg8X
2R4T2Z8UuhQXov+VtGQEEWZ3tQrvRbMwyCr0kJd4MIIsKYMs8XoMPJ/CrAei/TasxeOGbnEeDNvpZDPB0QEqVHp6M9jEbQAwamtq
7unyYT4p3EeNBbvTl9/6jmRwjxYduDEiqvRvlJIJVa9uDS0BFt69irM37Kc/4BbllW98QxZsuwWAxj2GpsIMw4H93LUNKgCJNbvB
+iWLvf88jrjalwUK3q4+gkTpTs0ob5fqBJfC9tO4c/QccHF/iWQ7k1hanUtkf2NNgJf98iqftIzg/wUNlHBWHYWX/Mozbr78Mz9Q
Lhjo8EO5BJDuypgXEteXSjj0racpNU/hrqpGX4jO8knvoKyP0DA9jUgQyF4HB+cwBe4CTzyOssHBNyzrFPwH2kxhcKhV/5qy8fwX
3hoIjk3pxY02hvOvPc2qzh3N1RRiRp8EuwYXCZ0TXVPEu2/YDREQhq2ygp2iMz7E90Yc9tkQ5O82n7zGVkPw9wEWE90bP5dT7QMd
plIm2NQQa8TIw4aNb3GTJPR5bcIJg7RJLSKAvBay6pU5UH92h6AZcPaABPByyA6eBCnxw+EQA3b9S5HEHEJfunmQNhu7sNfx9V5H
DlLf1JlwMpXScFHoJnsYkiV3/Bn5LYWnj7fSFRmQPYFUTOxiqVWMKFmiFV3V2yDdpSzUC3zeDZ1q2qhMSiIOHOLCkXQ1qLUMHGc+
MbID7z92QLN+O6lmqsSMYuyFQ6Pu+d79EFb4jBO/A+XPqNgJ6Zj/C/PmQrdvPEtzLD0FZ6m08zsgNj4sCdK6y+2dLyYyaSlTDfbO
6jI/GEnyo52TUaFxlQvlB4zF0qnF63u3T6As7ANjTqm0j2slzwBo7+sbxBJdyOdMHKmnIbBbLMLbWX9MUV1mKserNHGXCaEiVhN4
yR1zVpIYC5Dc01W7yS0vklD2tgVCp+0Puz4I8jOp/gjEjrXlB0vMg/K3wXgLuVP3/3IIUgIU4Pty46cBDZaQKEbLk72eXTAFm28j
JW78hkw9ZzoCVS21FPR8FpRNtXGMCzOvxQeS3KnrMHnUV21DJbL2v5qNoDgcAFB9qWwo1Tuff3LcF16ESb+j2criiIz6xXMYzl3r
LVIxbF6bTc/IibqOYcvM0BYxEV270+HHyUgW/PEjj57EYsNcb5unj1Ma+b6c2Cazsxe++8Ny/ze4DAhEnf4WrUpvR7XHvYj8qKbr
iErypqpWfGivA4W02K9MzfrtJaDME4zx4TelUqP0gmi8CLO3a1vo2QeCRudefopNqwVtvbnivX2FBO5+9UZUxR6KCDgX6s0RB0h8
PSFmtQE5LsJqHTwhgiJOiuAZvLb0aocdIvo9iCIim0HdT/XQawR7xlbDostMd/pI6NE4lxCYMbwdWySjwGasHPZfI8CT/DBkTpHt
mK0s2ShmSzs+TAEsD49rvasQGIrfl1nWN4gxFH5U+K7HaX09QpRjSutPnNh11SqaqLZ0YV1I5LOrihlKCQP4jdeYlgVwGEuyaiQ4
wBZ/0cNmbsIzGyp06IPRbcW37bigqgkEqk0WAYog4OsD5qGTUM8F4uKqgUsTWI0ML83E50JwohNu9pPZxM8mcy7IWnqR3+6ogsFb
cBwGWJkI+0o/iK8/UtZFPjAh7BjQrn2oO4Brc3NiuAKCrNwxSij0jujeKdkrxsvJOC4ZgA/CLnW8Rm2RRwOMVnVcJW/PXJwviAAK
2kc8IvdH5X17kcLhQicctAjfyMPEZtgFHcOwy3CPvAw1U96u5olykmpXPwvP1hvmlcg2u3sytyp/XSORir108tVoE3xRQeyH213k
pXYmpDey0OICQYx20zkfmkeE/qhNd7Kkgm2A8e4mz617Y6bW1y6N4tj/6EsKVsGfKzVt6RotVqtNU41I6lwsI8KyaL6tnF3ECYyQ
yWKTyqjnQNjSBZuObN+4kWLGOmH+cbW/AJCJrTBaYlIzT3/Px18/T43lnlbfoqVsv8YeJ+0CehM53L+NG7NY926yXKpa4qYlngTa
QqfENPbb/krQxjfZNoyIFZC4Vz1mcTTNZJ+PqwbjvyIY3HaFuUdv4NJmn21XkVwEUuJMlG6RrfUiDPraKE8sL34Ctw8NVW8Tt5e0
X1fBbQsASu9ATY1yxhes8ymmWSMICgvdiMVMzGHva41YfR8QU8IHdmp124e9gs/XuSMhmlJIpZW2KLHz3lQM/FDaDxP9WII/EIPs
hXZma3t99Gexe4qL/c7x4VnxAibaKzLVizLgn5LiIqvd8hycJNvkDrMAjzz1Ah4Pxk5eWvE+reINNFGoGW8RCi32jpLo5PgyGk3a
zfSNrRU6JTjXmcCux9GVJpLzKMzBBIB3ApTXE54kpM+TiyjnnhL8E/W362QnTHYEdkN+Vkg32WiH7Q+DS7696j1eggEr1WS4VCIC
1/yxqlmoXss5mLmFqaXPkbdsPA/gixKUrVPXZi8Z+YyEC7rBkucURM/QaAip+Mbum3EIvKZmuBr4NDEURpknE5Xy7GMNszapRF+G
P8s05XVV3NxZob7rv3+A6FAiteUNkMq4hd7AGv/lv9dpMqBqNZ6ANmBsMXtchDoaIy1/okDGZvgq11wGp1vK03fH+5KtduAH1K9F
ULeCBj2AvyDmHm+iVs3y9oHdmN7S711Z3uv0U+QWUld4aMNwWbZDrndrqNXlu+v9Vc0XIjKoNNC8nUN3McihdZT+uZED+b//30q/
UdlmB1Sg5b6iEFclaHDoNKKym6X4nPEEk8yjIQulQSSy3r8d5u4GExp0TdN6JEPp+8/Yfs+XK1DI1yPlL6RDtflRV6ZTcNZ2+wea
qgqQX2ZEyXvdAN20w6eLIGxxa2wWshKeUTqxQPw593UKqNc1YF2gTAQdv5ta6Vc2NP+3JDjip4tyJkr+aHys48vU9vL1B65qJ+z9
sK0Iq024+chCQGvrCiato/GbLIDBEZgX1otwWcD/KNZFR4CpjQ/d0Sa0agDy7fsv/6KAncGScYCjrUtOmvcpHwmF/e51OAE4kuqB
XNlIzbCiV05QPuk/vOTPfdPijhNAcyZVzeQqhW16tKSW0WvBWGsx7qPEldpG6Vq2D3nZw2p5xFdrujk+S7RaSZDeMTp5ecv4Fx3M
q5ZVAzdJW9sgL8ZskNtEaks10XCweamGaMeLsYbEfTR0IMnSnzOuuH56X5C5I3KpFtwXEXhqqnrfg/4AlUMLNYys8hKai7nrN/r1
QjiknRXK3Psg6YlgByr7MKqTl/ajqS8djDW0lBnxQ8FRBOobzeGWlhW2n2e+7tIL3BSBnOE2RjZx6IiJ+u46CbRZCLbp8Co6pgn1
nfvzsOHV5xIUsv7Ya4+7N7GcgsDs9y8Wz31b5/riV0P7Bf9HSuhnGKpYtqKfteSQ1ahXZURoidWRlHrbWj1fer7WbJ6eq4xByRop
n3X0pQ/PnamZIe07Az8J8RQUfreuKcypyCXKMqYZjo6PtTlrWSyRpykEhLGCnNv5gW6fhMyxaQ3a6fGsOXst3ebBpirRouSMmTlB
oiFQ/Vs6zkt3RXCmNF0lGxobaY4UhqSQfRxUMlCN6R9hhJ8d78RSz9TtY270P2xqSNBY2gBRU1+XXOTBrVcBl82dUYwfqNF9ytbO
LwRmCs6LGF7he+r+djG5ht8j0pQkgMF7Xi6dRUPzEj6my5Ka25cG6tg97iuaAvGplHaMIa6GhTfJWRziJZ/NO0ZyTSVez9f9xO3X
QxlNn5xSwHO+WDL+hh4xMhBSiJPBvbEnvITBlbXP4tm6sehO+AREA2SkK26o3Xjp0RtxHX3ktKVK2RqCIrEOuqheIA03w9sRowEf
mn89EpG3iNTl6B4rX5KWSqw/cEMANPOiAKsF5CttCxtsFd12gO/lUowFaLD+jE/0AWeacCZdKuu0CwErSfQX3Lp9tk/kFX7o+Q4U
BW+rtYEUiMvBziVXbdVpkRYGXu899Z8HJpLEPDoC+2sw1Er3pIdqEHUmK2LGRZ0/u7VsjpRHqlZYOMoPo+Ks2aWw5DodrFLzL9u9
PglNunV5SaVC9TTjVw04gAX+mds6DRC8eveYh6mM03VGwNkY8i0rczEHGjx16rDn3VJGRpRDzxNNGidoZcUQatET+SAPH0TPvou3
0GP3X0SBGAD60ll/1UqWyyj8cBLrd8R/uRbagjyGrAMCMH2g2gzUHsgWzWUhT1oKRqU3wenenET/f8Vvj9O+nnZ7khY/QlkSOBD3
vqV/gIhsp/KOkEHt9mdf2HG8mCI57mwjtUDl+zIUvUFtY2C4RFV9fS1EWcC1bCrxvcLaaKm5z3Z7AViMBzbZ0ezpuVuR+n+rUKsk
4QhU+aiUJNXDk0X+jb22WJylM/1kB2YxhzxclQk4Y6cO/yTduySGofEFctO+nR05UHg4TCMi7VHoC7PdEE/pvnhZEJA8XppfYy8e
WCmDh5rbGCnUknXVL5A/rRKVVbifishmgfgcyVsYT6gqBxtCpwudiHbEisNfHQW1V70IWzNVZXIymkEcCjXp7t92FG3I+2WjpNPD
aRvlgDNrlLzFsQsHSpsl50TZ/Zgi34RXdhVP+T3nq/+q2LvF8LHrHSaymLXEil8WWz9vx5uUYZWU5FzLMfQ7N8DHH1JY3+cOB2QX
YCaVhXssfRcr4+gJaWGZR/hfJH64GK7b44eMFbdkNFLFjOggJeItx+6ox9+eG1DbfhxqH7J2jUiK4sGYFX2hiUaLUW8R4KAALGZz
0W5pMjE18iSMyo7prldnXte7Zsjrrysv/C8EsHwhM83F3U8GOTccztd947zBPE9K1eLf2O9YNwtgPv8X1X1gzYqKOwHuj4/WQN88
EVKaAoDG4KwbkzXVD5kICnEMB0UL8Kja1i9bZuMe0lgwaJqqBE3PBoDh9xHMUr5LCt/ZoLVf179SA0UkmQzDTXSz2mbVkbgA651T
k0iOwsV6BWGA5JOIlFqHm2AZg0AFCQ17dEWlDkANke4W/Hc/youYPKYj1UCj6hzqu+ARaQwFikZhLCBhiRzx2wVMmgwdxREx14M1
npqFy1Z6oiY2Yd608loGVcBroDKvfYT8dTFi7iPFhT/wed3pxuXz2D0dkRdMlZ8AslWd95NUAdOD2axeLgH0Bjf3miSaAqZv1HN0
zgWjmogSa2focGV/0llxgWIJ0ZVV9oN6jK6wTuq0K7K3w1Ln/1537+33EMuoZdqxdRx5IlCPZM2xBHM/mzxvj/NFmU7lJ1LeSe6Q
aQHNqIBDPLfndSw9cjwg+GZSTWUYAFgDFDqW/5oahc5Jsp/sHzTuxBT8PhQfuV/v3J+eePjYVM+O0nDyhJnL/l7fqx0nivp/v5RJ
vCXdI7EnLFudeAZjFRQIWUAV4ZD8B2/fnYti0cIAYysoHme31+m/VniIu6HH8+cR/6I6MHkQ9ojBDsLvkc2889fo/c3a/IoxC+PC
0qsICco5SzfHEQDwi+Fturgzfa3wEdSxNBAvD3Sj/I9Q5VTP8iGRYmeM5UjbcbW6YHgmDO3yYZmYHIvFM1Nv0GQmpwqeQ1s0olm4
RbmTvGy+e7A8r3nRCZNE8iCXC88oKPcgpjqfmBXnNm0ZIZTwDLfDQe15srf4V/TPGGaZIFtYp+RzzSIqXBuJyg5CVQa47Xs9Knoz
9Dgxhv3Z4k7tOejlQZjM6oyzp0umhv9of0wN/QlbkCrmqHYk74XG51Qc0SVS02dO3lL3CT9Y0TOgRhlmTB2AD/CuJPU4wKILZe1v
iuvxb1zxuqS51qPvHU+kDwBIeAgM477zST7TISjJKGIQODXV+bew6ER6AS27IG8MYn4+UHWsYl6m2YDLP8BVIlpassr+wDfSk5w2
r0BfCe4spTFA0m6r6A6JbVmPBY4HZpDnUzWhSRUG2JLs++QyBJiWcL1yLFE2ZCQCgYfRpPMgDz680BlEM2GoaD/kLAa26VmwUWk5
V9aR/gvREY0891CzGosO74TpWS0qx0/d/7R1bTWe2IWD8D/+a7sHjtvMo3gM4V1KBvupJfiDiio00ZtbMvlCdet76/K5uzPiFR8l
AQNy5jeoyF5H1ZB7nawgeZnefw2ShxLHUas8a/YKHg8za0U2tA04q0WK9rgFT6z+3N22GjzDnOmeAr9d1DeV7aNa/fvqPdgrZMi6
dBWtQNM0xDwl9jsTC8QObubVJcV/VL2wimqwap3IcWh7Z7pzQe2e0kkfwTythji/HWq8uUZhITJZm6LkYpDeKm6Yk0IzQgVIWs9b
weD019h9w3tCEVTx78V8R6YlcyrADBEDZm7WNDdBbYtx5EWP/9K5/dxwiANlYqB9n5Od9okPLk7NtoII2/HhORSTsRm6OJa5tnKb
/Ptyv/B+r6DyZ0C6wvUULIGU3YdbH+JnI4IPTHtlKKlMRxw+O+5N4zEZWbesRsJk637dDqm7PX1XfH9BeUxrppaGIPFPVlZdnUre
qCRgVWjsg6bvJuGOq2K134QS6qSPYif3P13RDOn0wbt9syMeAWk9q42NZ2IbR6c8yL94uFansWhRZi47Vi2mHlftFWDqSZi/OKB2
aWi7dVESclWW2mXu5ykZ+QLp9BTcFDG0SYceY89t+DembAavsSdfJOpJcPrAYR56lYt33DhVrm9AtvlTDY09vgRljknBE6ftZtfL
c8uKroJIRMYHKY9lDKPMU15myoLQ7BpiaQVL0tr7bxpDtp/BbSYjq8cgRmvNocRrpHUrmONwVP/GYJKESMq63d4C1Hx/9+v/ZMcx
OYKaW7rlP155I3uTO5K9aMHZj8XfZz/yCJJ8o54FksqQiYqcHbK9d3CLMpzjViT3xLt+xS0CNoWzlc/oM9JFE7rvsx/IXYJSD0P6
H8GzVHrswrCOT1dRXhRkOr3p1Ksj99PpZwBWlRzUohFx7UW7SMMIOD/rsFrN1xSs4dTHPCwsWx73lkNL/JPUvJKCGRCNr1HDEjy5
Lj6vrM8SEov/IT6OZxyiWS8E6cAyBmHIMiQLLG1YKPGkdooUQu18BC8m5Qu0/3QRdij4FKu3/gPGWoFa9+xMMwWUO77G8vJ+2vPa
S8c6VFd04Yo6FVolze+6Y3bTgbdikWnLwhDga631NOyNnQ/nNEbBrq+JKMIz/1a3jxPuXwAgD7dX/y2NQOoVz3GA47s8AcOuty3Y
ydKqoHmLmGRDl0T3EWLR06obHXdFdvPPjKjc7kln7Djr21JUGbQF4Vqce+3vWA3OXp6BFAxZE1YaD5SYVMVZ+9w754Vb5Rq85RTX
GQh4v48/FsKU0YRDUz7oD3uRV4cfI2JmUs0HxF4skxkfsUWmBvJPhpC/g8hlPVfqWH0S4CuLsXCYXeTgHslOaTU/OpokZU5sZAaL
Z6uQTuinHS2CsAkwje0F5PIVvfBgUVEk78mFfvvtuKklEOT/P+6O1HFoeor8x1B1PFM4Toc5NJTuwDmTGKl7oe9nGwMWUCjzLZGt
zrQABLQxhhWv/f169BqsOA4C57DgbV3lfnloYICr30bglx2b3JPYBSBw8ES9xouI1kLwtfUN0w++PPfsqFcHHM8kV+zixP9WEV2T
dkicm7xrY4PUByIbRjrETYMNkvzthv/8OFkK+Vdi6PbhW2Qi6jswXQHLzwy1oY+TI9sMUNBSW1I7pb8Dcw+iqQpKH6XNeKOevmOl
2Io6p2MrWf2RoFQKp2PSYY2W1CkwkyrYW+HNb2cnSmDpVJFmQxNCmevk+uWKXBRMSwSLupqubhGAJYjZv4rrsPoOgKYwGJ2kLiex
a6mtq4CyDKaED2uJCFMfWlpVhu92n28ECoBDt45NeB1EMRG9/W8eANVVJNN94nl7Ub/4Z+DWcsM0pBjsLyk9NXh7hrfZtN/7Xg7r
tMo+b3WQvkk5qLnisO0XWEucWRrUP3EbR0vEjgiUzio+WZkLmzgaaY9KbmTrm30abyIoPkRKtUGMuPYYoxQvyCIG+4jrJa2QOOXS
U8u/kL83QSZc/ZK0pxm4REDi96IfCAzzh0YhVIe1GKI6ywA9C4jNQstBDgdReCEP6eNjJSHgkhMCJe4pZ+qnAXpYFeTTOHyFOh9b
2O0LjGzarD1CW5V6AmwDn60n5ewye3tWSr18vbC+9rOgLZ67wwVnoYyCRlWxbPDKl/Z1L4fnp7s5yWTpRNdYYmnMEOvO0AaKCZr4
Bra0huz5H+xiFf2hhvg33MdB819hyBpwJBCEZL6aL/FVW23wT65uZhBi2T3aBawIKl2N2C7oeugLU7bNXgIauh29HgFlmiCoQP7/
2SlvsO2Hx4C7TgNakbiHS/1nUVYdggycD1J+bOxwfofIKtuNlUNDoLxXSWQaLlBsjFcggnyESg0guhTFZPbQfPxb2/UI3XfxNUin
0U6L+NxJ0paYGVpSzhg1fNKkIdu/KzBgkHEJGpTdYMGXQvgAGin/eDPxy5KRnDyYFNKNx1LZsteQtNxTejAc5kE3+OlKoSwUicOb
EABlD1HpUDbNJoDR45O2MVlY1YtHpM9GrVhh6Q4z6/k4zfAwwy9WocGgsBHsYJDNhW33jvwBS7xLaIDpeFwInl1LNenNLuaIc+t2
4IlQIyGcsBUUDS97kp1PYIjYZMCoCVUa9q8xzQ36SEE5mMLuLfb+INhlDuk/bOnIGLH5zgkjmeNbWyIPR4QYegt/7v3Exg3RjOFT
TILIPzV5+WocDPVFBH5Cpzi3MGl/aCmQ9XXQCtp4OTxgKdG96Bu6GHN+LQgSNoEwA5FXEpzFkEvv/mysKByzz8S5uStoQcES+3aP
PBhXY8GAOrfOjc1FSbQScvP6wm5S9FLzD9dMD6Q01wtUWm0iGW3/+NfgKbL8+Y6IGD/p8RO07Nb8VuWkV+ylvyHJ+oUOLxtsL9FU
Sj3n/9Fm9uOPsUvNZzBGTOPHRXJmWjKiBUN8j35MQg4S9T/OK+NZZLazuqIshwylq8F0jg7e/8OdkS7PowopTvg1wS+2tVtRtLoQ
Y4cJhIWFPlPC74ZMIoHIjf3Md5aIJCDsFldprxNbmAT1Ir11BknR7ox4yBCaqEyYpToUr0QEzN6NMyV86hJ197oZvdl42jKoTADf
c6LyqwEhPkEWTEbGcI0Pp9pGnhJqV1gWGn+yTkb8OLbyVt8+jFdYvPKac9zrue8doC4DNwBVz1Hso30kZMGb4/w5ZRJEiKt5SYr0
IN4RlkdQ4Vy3BZVgvBTCx3KSOZJT67Y7MJkNqNAaewC4pTYWx1JOHpQDOuPjoFfbhWUQWAzpBk9oYTIsOOvvGbOeGTSdFsvD0hjN
N09RKQ8QVtEElblj5uroDuk7VCkFx1RQYzidaZS8Qwexm67OaDw7txyi/8PWNvGZGTnr7EtZUPmmCXlc73SujJJQ9uBQLkHbkHxg
6hB1Q4DjrWNeLTSwbs1F7VvB2nS4ydrrIhcprr8Wlk8Jgtsxv8NvUJzCFxnrNDqVQH4qTwlKoN1uDPcn9eyUnkpndUrLnfC0hDkh
Ul6Cn5Hi1GMm5ItDUrSNrj3RD4NxZN1IHoiTvAQ1oHo1VDePQZr0ZsgLLff04cQ7fT4uHUH24jBIDKlCio99N77iww/w7okGhJ0S
HgD447kjY710OfpBRRogxxwrwguJ5JteMxULRrtwo9mfLfwA1Dl9x48d7pEJk/XBAA21vxd1HbKqSjFGVmDQMTwoFCnrG9sWIgRq
19mAgz9CRHxU5hWQOg0L33CbzkKUFi1KYRNYWNQTKGK3s075mjMEztd38FwVSuVx4XsTKEPEXdpUCTfkzWzwkqMnsykDMyhtzoGo
ATxpBubf6QNFMfJue1eR6cJaDuZJFGNDeGCHhePeqzSQqPABzzt7oceBC/PfdSuEEr4iwDWiNJwcNoL1WsHKhDOHriXE0/5voAH7
xDKljNgM9YYQ/xNWhKtEgITMdf3Igm9qaDod5sxC/noXpwm2afq1/94tyH89I61awW9hTyl8+44uXbZn0JKbdc5qzcAPplxQ0cTv
A1QVzINvHCJNIrGQ+pPlNd/6a8bdoNH4ZbuH9GFEEFIXUIY6WcSovDrf8/tHi44ne+5g/LUB2/qlAQOerPVPyFYrfNwOfDvohRxW
u5T3s6YphBjzj2bgEFiaQPl19MBiTCXY/Vs7vxZaFKeSTStoPmjrFbrarrmZXVEtTAbBwcaPW38OpYZ9iMNjtAEVJDbBZhsB/xKl
v/5p3EIU6zLtFXm7pcE+2pdXgaskqg56wnn+41sDuu7tjZkxyH8b8vf9rAsSoB2F3ZsV7Hv52tIrvBK+uYK/ImG3BSivX3bUVWot
MjJJw9uF/Q1gE1vWv8lPTqDewLrlcx1N3lC1H7HVhPz042ni0Xb2JOHsEfVSGkZr49StAAkayIdj2LmRF/jR1CTtNX61esJvJJ1h
kSBdfENaRDc01SwPsik1WfRkpLDwclWZGjmIvajL2xAwFULlK9ZJURAOz4Ag7GHMkM8hUFdFQ2KYEHq0DfUgghVyfk1AFpDIzGdV
ywIgMPd9xvKAdoNqPJde0R2f0Jp3gkw6QYOItmU05gsSmjcnl/FyZjtOl8MfVuFxxJr9Sp6sFRk1hQZMaITFzaV7AC3rBgwsIC7F
fKvZzQcbTTMuYLy5Ojd9t8YfXMYMFkBwG2t9JV9SPibO0vB5qu0ZyF7Ib820XBMxiJEkdQ4mpOp84wGc39S9VYXIF6AKvdincDVH
nJRg2vRm6LX34eJpSeaMMlK6sjwN+kVT+HXzm1fY+54/B+NAXJPsMl7jGpx5iu8FjDtfmQep0hcbQrTzglLUF3IYRCGEYB11ffMp
tSoYYX5bC4adp4D1IVXiWHmezoEoDxI8dcWz6OilhQ28u/HjbyO1bCtAOvgb/B2K/xPSBzNbrB6TtAx11w1hPMs6GW1iA+P6CyhF
DcKLULhaQHiGSmCDnwtl9qFFyPXg6O9X+k15FFg0RCk+6JmsoGaY2cZ6gOL8Hl3oTHO3K4xkAzYME0iX4WtZcs9fgWvFVcGVWawp
pfce64YDBVP4LlMch8hJeURSXNNN0DgtV4QzkkNj+q5DDUz/ZLc7/z+mJarJGRvcA4eZq0CsXABhNMJG9Smtqz2kZAIPl9RRwHf3
sMp3ODH2uQKdQIwJqz6EQiY2nW9T4+NjooEDwHjoKp+13E3FIJ1zaSCgb0bMypUUgHQ675/lNMvhjxY6BNIS7NfgeYCLzpTVZhHa
Gbu8ZhPaJ+T0GWPD3OZRW+18tFd1olqNN/3w6Ddr9IpagsyFDEZpz2g7n/6MQGxNio/toLsfcFWYNaOzWhdDWE3iR+kcUiT7Lm3I
AU9WeOfghy0Pmydhg4ue2rmQ7M5c90ZqCLE0jFnCMw71NBF2aG4duuZ3rah+oNm0va+8plA7ZUxv9v6XX/xZpGWCssuwMTSlbmgX
OL2TOEK741A1HZF/ugLZCqzueN3putbZEoSk+j4vdOCSCjZENONrJ2ARX8zAoAV7kjsyu+2E6nb65ohSLb0zgI8JkUw/VcYp3xLv
7dkLHh9UBsnmYOawdlgu0vQeLLIp97QJLTqZcrc/yLZAeeHYQjz1oS4jgzhQI+V+pKQYFocx2Cvtoe7FqzWf1S420hRnm77RaoUg
37Ku7XrHzoOY3TMzMfvozaor4aUCW88jEaphVXICM+akmmHO+lIdN/uNQ3a211eC93+P8B7BIDpUPUJF5PfxdusgEm51IPV0A839
0T4s7yCT+Ekw8r0G8h4nel98WngRsbkkuQjy9zX5LoqB9hRG+PKIzzMiR3AUn0WuOFXINhh8sZeTEPZREy3/lr9HZDhbYAkRUxZS
AXMfv2PEq2Uxuyvvmf5tqsOczjAW3gKQRbFunCzUAcn/Xv3WjSurVOGzEp6YuxJxZR6A5GOqImNoWLCTKqDk0icUL2h9AZwoakt9
Cc+FTuZ0eGJtnioXUTdByXOA7btX9JhZx9Fx5E0yo3JpeOid378Sav+4PorxvcGghUule/k3SaIFUUsnm6yLvP3iS89M68tZuc1s
zp3eX0KKXM1XiKToCTqLqyDuv9aoHbIvTsBV8gLex2OxCKhrbE6gXFnYxBruRjk7nenXF8SbK/ucrpTocfLctdyQ6g0KdXLxhQ3N
OtJek9/0KrEQpkqpfoSPMJwvdjyLbtjff/wQ9YsIuji9ZHdHd1i9Ca49kv2sPrwjQdiQ4W2OxXwJhL1hmhtYPBuNl3IU9e6syld0
mGB8V5HijR+ADFBMV1Q7rkgENTXOsSH0ZogwjmvuVr9AH/ziCx65Z6QTEzbGf47ghrQ74riBOi5QBtGrWqkyMkmEbN20HKHIm8Yn
OEKxHY4eltenZxglmByF4r12cl7RzT4AmV4NUMq+/lAoOs2JueBqiq/HkIrsRruQGiJKm7uLAmXpcZsTxaVhu9d4Rqt0lyWMshRz
AlwMT0j1LjrRjpAMzJl76OOBXBKEycSJTzWRE4KbgV2KJ3FguqfRCeNAaET1Hz/M3HluDad8sPxKohDAxNfKod2eqWmTFcLLsRkO
s+vsMZxRBLlPRQKjNfMpIfrBL47eBRy0yHz3RcTFW1R3praDJOZDm+ZFfJmmNBQ3K0Qxelx7Mf9niazi3n4FJJrGQX2mOjVQeWNW
2uD1HjWlZSlpXYYTZTqJPWVLf4uRNM+LJThbqkdfy/yiuM8Mb8RFlx6gIuANkG+iOdZBTVFupvfpN7eAP5Yf/NNKMy/FQK6BjuDZ
xNsFySX8VeNDnU2I/u46Un2bkLN9X24vx7C1M3FVFFT6hteL1GG+zmjJew95nDqmByd+kEAF2S/0tky3WxHfaC/3vAH0mHigJQHn
rU2DdOnGX7zLRmoXoiaHu6/wFUpv32e2xyOLM2nQb2TP0+lxZfnaUZK4Za/6GzHtF/GphD1G85e6VleS5Lt+Hbj0RP2nq3qiEGqG
PkDR9WxCHOOj8+6QokDaBaHLpuCZh+ccfLSH8vasXIoLB6AUnFDf/Mie6T5/b78E0/fXolFst1tjorXtLCkyL2ryLjRF/OU0ErH6
fxozlv9XkXDeNOtIBU5iabwxtgWQThcYJUtzq+VSHwnF3UB4lEP8wUghKEcZz7QxEKpr3E10rIhOu/oQmvyZuJVT6Ep6tFWn/SIj
Jf0YZBLE0vBhgpQf6fIbrje1uId2+9zf+KOmTmxn3My5f7G1/+VU6jzYfBfdckW8C1Px+8E8VMSUzPG7eHCH0tUnUvea8ItJynzh
1bgvsk+Zvh2tBA004blP6PIe+s9kcXg82DHo+qt4PX9BurcxtzrnY9cceXG3ixheqd4Q008P1Qpb/OtCLEpOIko7GYaf4HO0CJWm
RoU0UW3Jm0SoWDomuHxP9i9J4/Lu60xuZOgcTUaisLmjZkLP1wEfBAa/xnOCZ1yvoIALiWOT96/ZvEvrDuphhfTEf2J1XnxbhXjn
SminS/qSamx3Mgi6LuMH6Ewoo+5Y78g/mS8X3OVDayTa+1rVYRZaQUxh+i+ib4Yxtm1TEn0ic2TQuo5ABQLxjhySEy5o3ojla6aT
iCg1VdzgQfqAG1pffjppjKEsbd+P/QZaqPJJq2ECus3GQ0XItuu6ZW2+sNAg+XlOyIeBt5j0xYityag6aoEewUQpJiwUIVfUjJVG
Sxuw91vJtUW60h1OWPSodVCAjS7NLo4OfoOvDTPPJ058u9aq1aX5YM6Py94UuIhN/92KdGoe3ibJpjSkfp0MhidpAN6LlKRT6ogi
vX6i9v1iHtiSDFtXoWw+cjT1Zujj7WwSUETNLLYHCpWZmJw9Zl/NWWYvZRudaeJ3grWIRUaGY6Fl1fRURDPIFIB4z5G2WF1WiN0C
nrLFvJ7yDX8SPR9dsVdvZW9MYex1nYhENQzNweYJq+LZJdjRrClV9xn8XdzcHbt6VW85GFT9VHDz+PpwW1aPwturJwnTI0Gi+Hdg
z0XPcEgMuWVQ3LfujP+5MwX0LaC1QM6c+XxotyhTctf0mW8+xml91CtH1qgSmJySIod3N1Fj8ow/u9ej2VV2MQ0hzg2rZJ5sLFZk
w4YM5hpoZrEzEcIemXQ3ZCbBU0VX0u0T2Mi2/N/kE9Gp4s5bNZzTRJDbucXWmy7o5qkzImJEBl/AtHIKAjvHunUeIgrGG+YePwV0
s0XPU9WKaM1pZ815cTf2cwtcHBNJVA0nFg7o/JZN6Kmqh4mIEDGUy528ZE0gaKGDv1vygcLaan5eu8IO0pgwTrH0CktB95Wjapl9
JtnzwyUxHvoV3apFLQnbWb82q9NOYPQsorKXv1BugEIDc9YbfJmTz9k1FmwbzMJqJ+0Ih0h39xDw0j3yompEt0H4V+zdqzVQn7mO
wxbaFpc+n9laTf/3xCKadvVeJB/h8B2RxZVnGV6mPRvNNKoRg3XqQfr6wtlIqoo+BY/EBFFWG3kLIWO+9vqnWoICvoabu3CrcN7P
cyYnbG6DgqrvbnHb/lnB7vlVZAxhltVwawIGDImvfwVesJXA1t8HkbLZ8G1EoVOWzgCuRqvH5NijcpO7XGO0tjSuLbzgnl2OQmK9
GfCbkVF8n14I+7m/VqXTWtyzivoHBy7yaEIYT5gLoBFidlbptPF9zVDqxQFYY4jgMbWcWlPNvA4gr5cXkGxZGeqYNbm7E/sxNVBL
9UqW0tAm4SOz40W6vfweJJyjZ95+3RWtUf4GECoF3x03lYnNZvs83yJgUTebIvfMPmPYm2QXPApVnQLgK5VE18bbZh0b82k756jn
X/xULEiSbN0Yh5rmU7ZagCMYbuiHmYuhO45ZWG9qEF+ZXKDQU9jlG8UM+uXB61IsAzuWMAZENSgWUFLRk+gW/fmnpe4MZ7w/uDJD
CnFuj2ECUdewjrR4ZgDjF5jdUV1Xj2jj46CtroKZRz9MHsjdCUPG9Jz5Iaw3/ShVGQHJ7Af1ipIC+8SY6eCyItCKcBDqtgmVeDKY
q6E2tQLdjL9hJ73218uTAIqZU8dLW6uWNItYZUx2L796qfysWUhOx6mR0jhDlIMf5y2M+6CFfc0BJdViDt4XsYKHuETJOO7SOjz0
FObiBgv/fgtbOCdmHmt0TKau32iIGAbHOyDTcrpJkfv4Ap9Nb4X9MHaDEEIMhHbkGdBNsKTuig/bAA8BwTSsahNYTTd4G3ljJHC9
280AOlIaSTbQcdNZX0mvZcO5SjrMfyJ0c8R1Mz72qTEYmdMk45JPkWEyUfoSyW8uv2R4HlZkiarfLd1V0T88MGn73Ye7DcEAfDla
RVN4LcZ2XZ24aOxfGcVOY3xMjnpQZpvMViv4AukQIS4tnGfdDaFRjcnkZi7MYK5eB9WDvVeaovwc+fUavO7gQRoFtyxGSBf6mV6v
YVi9aLY0FANNKrhMeXVxhckZYYLA2DmlRj2LBDEADlkwnXW/n/ffuDu8rDAaafTXUn37HP0cKJR3TGK+1AVmT9BNIJo8WjPcodNr
AwDEBDP+kWG0bTqT8glGDGbrCaqgaDQneYCs4NZytT+g+czOo9Pj//DqtcRQFwp031fOQaBdgV0jiFgNjF/x4wBC+CYYDShatokd
HLS/7kxgmDunnml/ejkIE89/sE4biG6LEEOONgf1HhvOXpnlZylFmZ1nOSVC6fM9kPceLGONC6CsUkIcS6e4sGWDpFLSkczaU4Qv
c1KYzOIGcKAG+MUMPy/qzp6cyMgJ1KOHlgmGGbKnnsm10X33PkcMQ1OfAagPoXWgV1PEfdJSxzbzeTA39IP0upYjsiHXtWa4HeD5
8XZaPVWDU7LP8e9AQIuU8MAxoK6VpsJ1764qvcNtQ16xRrJs0jsZZR+6IJZxn7gNTfDnzeXzYA2/LszxjDtRmTP8Q5hKwTlNXfse
KqkTWaxQar8EqUKJg5Xd5g6qY8yb2fhv6jpYmjGpNAYAQQ+DIn6lBCly7rXGMOPuF2fE02WTtGNMo94lP/68stAUsTEHDinl2o9a
BPcchhgpYifRc7GnBzx6wFdFxcAObmLmgv84DDtIu0XoYkKCepyvw7uY6m4dDKukQAVM7gYBFOM4UTfnpfnvYV0TsuRnoXm6WlLJ
bGYQTvN1HUBErepHW0pdGiAaMInaU2wmRvGJ1RE9yE2oargPXhEnpyllHNEzEndrFtnf73ejThfsFxnZ65IworP5auZpjHRQXAll
TuR3DYK6i3sJikO9ZMdML37by+w32pzOJ5mvJoC2tSg2g5ZfbCJHoxLQ2LC9VcFbrPcApxJ2pxQMi2/mredoYQr033E6H/Qb/DzQ
dMziD5B5kvEs/yhlq8mkXb9x+eObAoykple7eVRBrhEz5dsjmCTDhQTUajj2QGwW0KAg+8o5qUEdcn5RGR2Ox4tyJ29xA4RKGKbt
e8k20yCWidh4dyN1MfJDQ55fhLsuuyygGaUsi4BO18f3wQJJ654ceTf/dAu9meCH/RiLat4LuhWyq2D83Jg5IrPBQc3aLgIrBWHm
ld+BX5YL/L18kJzHXHFwt7DGnahmxmyoJCKhi/MR0M295ZJe0RX8aTZZlsEtLzgK9flgmT1HGbNXme4l4eg7ryb54hc9haC9bR70
DJZgt0erne5DSoeEy5OMFaAuFEXt1+UqPVq8q8MHGsixobcR/mFibZSRWe1nGOwb+YxT1pBYaVTKfMuSGQp9eqgzFnXb1CCKH1gX
CyBjXm68/RJYvOj+vECSBUy9ytvQ98E6mmEk94vxmETQ16UhNZ5MBaWYINyDHkD71cdQOocPTfJ+7/8i1WEw3okzv0hxzd4zVwxn
hzrg4GF5R2Q3qojfWuBzT3QHHAQZG1LvDPZUdOd1J4a35e2zmcUmEbT1nRg1T+YvDQMqTrBWVdcunAmD0oTG4ni2XvA2up6YjTIr
EoWi7l10wO8Ww/qh/OFyilYyyC2ZNcGTnixjL4USopmZgFaB5y3mpkcfmrWmNvLTO6Hq6E6z9cTVKAyCqEsLrqBl+rM2lsh4jy41
k9QhoiLe+C91zWkJeJj3reMW2Qetk0b8u1Eh1i1Sl9TLgVQsbb10qOXY87yxd3PAClCJWo7ZhAoFQFCfWkcBQqLr9bqqXLcqghMB
QUQvatgTVzn3XNyn1Nt3OASfTYj1mBjcvno6oty0Oc5vG6Vkv1rrb1NdsFhF+OpE3zOGwrnKb2Us4/rKMDXuzRlw9nRwvW8EDCkA
Xrp+f3RVulnFoqQUUtFxbJa3Z7FCczKUptVzmUXwfA6MrwfToosLx1q05CogfQrR250MrHfsLzgOfvvFPmHvLxVkd8RCD47EKdGW
YmLOcIh+p9pQkKew2I1WsjdZpEJkOeNwDFHX47QnK9UPvPsHUbCZey8VMt949zLGxUsGhHaaP/uKLIn/I5lc+pV3vleuOF2qnigs
9wxq1JWYB1090B5/4T6IKFHA0DL5sVXeC8S1fQYUHebFjfox6C47VqKdFoiOiJ1l3udBPBq44BJgydPeFL3CzWwqriwDTUTUp7PC
M43Myjb+Kcnoui3QssN2eN7vlEnRPujjVrF4Jzqw58mPGs0K7b6U3G3vwS3gm5iZskb+IOz8i8DQbIaOoB/LVnRMM2e3hpdp2AM5
9B2DCop+zQ/FWmaGTkKJRA9ntHJe6ABha5jN/9MQk2cn+iJNS7pKuN4cFzDiOmS2MpQq2lGdRa9ycd0ULgsTTTY5xo8jSv/oIeLI
115lw8B4Mev7akezmvazgp6ma1n1GPu2oF7n8Q7CpGSL9Kl3T1ZIzb9IFdT1BlWi6s55ysI8U4Bja5W/DjV2h4xEG1F7jmxt6it8
emkgiJ2meD7mehwoDGgYFErz4OutN9Hk7mvKrJeK7WPG7mrEGxdha8Ae68hK67osAoVzReHLYTPci3uBVeOrNc93aIyWZ1L5P4AE
Kb4t6VK9MtlSiwS27UB8jPE/MGhMUhcBeQNMNBlxSySDhws6E2TcShMxU/LXLr0hmVYDY6drjavOzfx7S4gca3aIwWnt5bTNqY1H
Sl941FL1PHdDF8HEbSKqQ0LPDogF6G3P1h6f+qhGy3nvOcnqiGNfKGD4uHZHbx+42pxXOdZ70d3O+dwipi5F3AtLkiOi2lznumpY
r2R6Fs7gya2RCGi2lYYlkRjZOObdroGXBolh3l3Z/XI6pcPjjScfSinNrptlaxlnVg8niI2+dXxf7GoLuEiTRc8WY0buoM8qcYGU
0KLds6HYJCAgxYu+EvuAzV929AsPl1NZ62u2wEUkxkcNAG1g9N0qrh4547S1aWKYPuECShre0tQd6SbWfh7SH7T4tqHSVvp9H6QI
KpYkBDacou2IB8DrXxVeiekrIo/iFCGybbdnqPim6ycbiNIn47xlsRrX/gtJhTLfxy+iVYAkse4/NgNSZ8K3PqkRnAVdtSXR+Duo
NSzESiArEtj5heKH7fl99nIqfcEfJAB93o7228eXmgdaNbHK3IxqTYOxrMBKAjbFSC4G18GuNcxrlY820LS9OGwVVA/moA8qv6SZ
ERZPW2hdU8dpCyaR1JsFeOvR4AbnJA6cGc0U76ZQxKMpspK27k64X4jya8E0P1zY1Vs5rcTptFlnTPi9agX4p0B1CjUoLZYz5k+m
CUGzscQ9SOYsEyyIg4HVsiV5QxVkbZ6C1w7hVUCCV0sn6ltArF5iunFBRZuyfEqr1krtVE+xXDzmF+vTPdEZ/6HhW9nIgD/lbT48
lotVt6wClyPnPKnRTz3r3IO+VqGOKcmPCAB3sYMj/9lvWl2qQ8b2SYGmg08CAZ0IkojPLUaPnAXaepXysVtA/0IPOt3PQavUtet6
02amSW1FEqY2TtcV0phOE6AT5eaNbLBBIA+umw8l6APwK6XhbMIQ8kGW4z4mmTUVWsaSot5Y0E3uGLUFyhHGFcNj0icxiB9vcMOU
7vCQcMd42dLJM3vK100e2QgMbcYeuEoq/Gve7FIALoPo6ztQA+qH/RsO/z1IjPAECNXvhGWH3K35p3FbLQFU4Xm+coFBnKL4hJTW
nk7502kgPAerbUvg+u+HNrki8epBSLj6IaZNpozfCOjcxfzTy+TagqKJtV8vUdmxjuYEGPCpexWyae9fNiAtvoZasUdliN8RUAhm
EPEBR0h+1ye6gIMXfC3XNQ9MfK9fWs3nb8mQSEzAXAtFwkRjIPQA3eyBSJvqP/ktAJVb0ZAzIe0eHWXG4Gpql08bmHw80jS8a4mS
2h70DqKYZEtG1yE/ESA07dz/U0pmP58ofR8g1ASTiRJav8W8BTEMkyiUrXvFVJOG7A3/fWstcTrDP4GunK2DvzbYmTOfyjGVsMVO
Y85ExnYlT6AYECHhhr3zvFJozgNVD1ZP3ug4ATgUp2muJGNOkCLK37j3IsfWT0JAneP6Ipuy6QZetwFB0U/wpcjTE9Nf2HzBvu0x
UssGZMVcGVq0tVb6kyNgYDWLtgUirLvgV5k6SHI++GV3Mwfh0Wl8EHATHlhBgSkIeUWTKmyQFnkhd+6+8sZA5w+pEMB7KPZ+Lsgx
jcX6kMov4X+XGBpAER07hcuf+1hAeCUMxYUJSJB3Tf1qEu5U0RCcAMSiDXwxRSpJ459zhKDxsDFUtOXug5qdGKdrhGiZ/Kkz7ak4
Jog4KF2KZuNuR9qA8BAyQpp2gWs66sgxxpgn25UqfqFGAePtOourl80vftvyR4m7zMQsm7oFSLZ4/wCtXgKUzN0KREJ6zR3ZaCkS
FZkRAPDex3SFoeadPWd//0FVvUj4/ezK/rtP1064YMX65UxolkHu6fLxszUsmDbfp/KRzYqruFPPlV9yKApJAL5m2yClIpOensw7
8THFzsIlzXmK7aMPuvuqQfoY0W85xhBMlpA9fRuauMyYpewkELG1k1mPuBHOQpNoPH3y6/eEq54s6QXLUmNPqvDv84uAO9ZEnwy9
+NeFm1UxpsksEd/av/FSV6S4KSlflV2oIDs2l/LCjRsng8YwD26u8I1kIBgmNvNA8m3QpNwMW7+l9s/vE0QU2Tu+QiQpTVpZZMe0
2UU2Uv5vGX1Ollvb0Mv9zZ2Rfe1HHoafgJcf5aGzNjGQf0//wdd+BMeruf2DpngAnPibq8ihUkeyOt++HD2b1p4hZODPqx2lXLJ3
Y9+WCg0v495N+R6LiL8m9UzgxaVHkAzulGGg11lvrFTmH7Q3vj+CGEpA3T+NZcghyuiWSA+ty6k4JzOCQDmlSw/9PoUusTDUgyu9
4fhgzrywpShtRg/OLWtgYzUq+qOakevXDLqUo2UFwlWmi5yI9+HkZFt7ByAsicsJNsrioCYprvwdjwZXzPgFYFeA2az4SOAeQj1p
FKQ8hiPJdyBaDfK5DlQNllQuRC5uALHBcl/dxaAQB8w6Lgni3B0UkVQit8eelIC9QCT5fXpL2DqMiXBb1esAdxkfH95saxFsA0Bb
9J+pIs7nKMaICepB6nX0jQZZ4jQLfT5np5UAN842v398+Gkwyc7JUyJ1xH4Yj3hagB7HuVs/VCJIIozcFIcRFZGMg9vDlO7qdi73
sTzYDCPOPhX2hcDGDStarUsBW0w9DQlIxvi0pJOajTc6MT4kXtjd3p2B/oIAB2NAMIeKyvNlDVZlm7fODm5Xzo8K0t090Yfx8gDD
ch7Mh7A+lM1twu9PmSKbUtu+fyshrcQ+I3dBGC7FIc7NVaz+YTgtjVYsC2oz8HfUoYUjq7V4ISFZfxoNg9pYp8BR1vFYN1yDMIGq
5DdGoCuvUBAEUyIIiXO5kFj0C4Ahl55trNiZMtc3AAQTNa4HIQQ2W04NPVZgd6GzOvVAZJ8eHX+TJ1OtLGCBE54Ms0vOzhuxMpEf
mIiktTC5ZtiqtEBEApOTdoyB8b7rJw7njqN4tUQyM/yjPgi8lj3wR4g4KJ7AXsPIU+WmWP8cksApKqYWev/BIqNdzR9s//ykClvY
AOdlQLa6FywvWfbq5P4Kl6zgdCSzsgBOM9OMfb939LvhR64qZkeocE9WAuLLFkILF+9H4luDyzvb/nBcoVqZvYN5lHPbGhP+wxpX
6F3uwVCr7ccFnexzI2puqtK8Y3wpb1ZS7Un/OhQ+dRnNFpiBcslbeViBE6GWFcmN+qc9lDkFmaDA52esYXHciO/ire+WnXhyMso7
g0oQLLqwx/8VV1qieLPF2szsgsXBYSY3fpl50cjKEJUj0qYCuaf6zAce8GMAljNymEaS86l08WndIvws2QB4tdAmYfrP+Df9LX5P
WQgBaanCM9ka2AmIYd5wPaMCQWMpKUMNqT43cMCFkJlIf5Lhh0qYBdK6NM7kLJV1yVKL2AQmhmSQnChiWsQ7J3yuDWmPAsw2uQ2x
0o9EEklX2JBpiY21FcT2NlPeFRtwfy3EP0eSgybixa6KIiXxzUQrZEPzHsDnogzmRttaB05vWp1cPIGf/FbpNCdTUxCCyUJrKGUU
Xdd3sCswcGe+bX4c/thb7rr/weRTQl8hHR/LxCiJf9KYsNcUn0Ai+zUBtqD17OM7rnE4Sji4iRBa3CZOyYY9gkgbXeIoiNAfLWl0
GFPmkzUec/YUoCom536p5mADygV2VRMlFTJmBTvb3dAiqYgPE39qHTmLkqvHH5viyqD/op61usyMVdaX7dQ15oEEzUYyPfooa43p
hXC2mkaaWeq70drGX8RjpIWLc5RP/e7lpYY/cVJ7ClTgAh20ly8y6sc3bec4On34FRzh8Wjt/VGCLOgiD0k/5rqWCXHHXm0I1lNh
44WQi4J/bPnv4MG7+mCunczOkHFAD6xH+OyiNwH1c8dJ8qMLqL1qKOHBN+D75omgzYb+wn0NhItzsr9Jp8DqSH4ccg1b17XCEjdA
RVGHvN/pCIS4w2IZDRLwMzhm1vrWWbvqa9RdO8Y19roCAY3R9Xlcg55FVdv6YmmxlCZVJOt26IYbzr5uWvAi/4P4593q1RiYnnZc
k5HLswSefTa0Mf+/CGnsOaQYKocBCfwZQjFA0vEIBtEXRNSMU2zuK8VV1qTAhlWMTlYvzHaiOZO3JIz3ShCY9g1KJMCpCfIPWdxd
4oAudXWXvVpgw6L1wNQqNLKVdT6unFSDxxFmHN65Jl04hn3gPJZjjOBGgSErneWpFhqubWjxOB2NBTlvQjV8v0N3QExuKyGRMRua
XeKemP9k9+5bbnCrhthbl4HkDkVHchkCgJJzAB+Dp4dmab0tVRsyiTtxsxjBdvnsApTsAUNWU8rkN0eUtBRaF/AsTt03D33Cpf2s
N+gUqySV44f2jNTWtEsz5vSOOAeSj9aqdf+yZm3z78pNtYdtw0lUaJ52RPisoI/+GTvB+PVwDalrgNhhfq9zKSw9Y3LIeLFmxy2R
13klPL9xyJJCfgNqGnXBMufRYEwlOKry3OPdEGneDYcC5WtZfXAFxbZHbO294OSk3pjZ/R1qhfsE0c5tWr0G9USUukC0KBnBpL5W
EQiLiPeon6xccPYcsSl8JPnvNmNqn3ZzjmPagVrzkqJHlpkiELL4RWMhXjKEDHPX6MxIlDSy+m80dfh5ZNGMcofmiKAG22JxxWmG
aZNddJnbonLXXfBBcQmuBqr1gdAsxl54gtEvZUl2cMvxiHjVRI447c9/Qsozt2jAmy1YJswN2eG40Kyc7BlBPUw7pMDVCfHIcu+Q
iFkMFxXk6dGRWbtN13ovZ8UyS+LR+51lBb1smveFHcdK5vVCMabU5fxM4kFXNmyk9fmSw6ZDNxrT8H5idwa0F4HJLPfVCIMYen3k
UeByUCx2m5fng1YjmNwoFCXIBfL72hsDo0xNFhoVFV+D+SmNOeK2ksshiakLmv20XWvwE5Y6kVpfB3YNld/J5iHyGgjNJ4Z1bjso
QlkyEb3qSPhT2iTyMtX2MbX7pdG9czw39TQPnLNTO/i1YnMNLMME6iCwvBzTnjY7CdbwaDmgVbvManwtq1Po6ds06hCUYLamQWuh
58gWCHjZ2PXEJ8PGV5/kChJ4j803sMr63iYe2CcUapstZUdZasu488vPlpMM3P4GrUa28oXXiE+dKSpkznVqqTB7O1NZ6Ltpno0+
ddcJWqdlCzUNVWGnClpGKKZPqMb/oEA7etxSpopPT8SGCmh3jGacNxV15OJGebfREKwCBoPFscZx+MJ1FbP3pP505pMTKjlIuG25
/W/IOgn6rEy0ZeojQ1QMYgmwwAIFf2DERQJz51TUdt1IgNl9Y2bgnTeHdjBB0z2oVRPaZWE/Mln4lV1ZOAEg7iQfjd0c0uPiY94R
aiY6QmuhsxR3cGCm70VuYJbh+BRRKZQgMb4XvWw0c6y6inJSq02eI4hDwJJyF8Inho5EQXgrnw22JueZo37MJGzMDHO/yU2+fY1U
bhmneISCl/DT+2vAHOxo4MS1CHIThcwt2b+MTJi3WuJScmlk9U0ee2p/0sq34SJmHi5QuWWisj4tFHg28nB8pvlcRTb8kamG4LNv
0soqE4ARHwzG5DPTHhGcimJoOyxHaz2j65mAdE2oVMMcZeIUaX1pl04bD+Z/LEw3xFn0QL7RCZG+OExN1IoyXuvn51khLihp6K0z
FRPXGKQz0SuHZXf9OFB/t67JDGXwxnLOqE3vIHGovcT0SDQfmVUATwTWEYYtN37kN91/WYHJZp5OHOFqBio/lNBbeEBw70MCzc7D
HwKmHRcKKNcqK412y0jMTbHbJ3aIxixHToIGaIjbTHNCj9rnzoG6noO0caEg/CR1RNFgkGP42ERuAmjzIbSJJ/80HMG7G25PvgIm
ASSJ9ICz98y29HNRDJpyzHvzGC5MvSCQWpqI6BfudQk7OnvAD23OjOQ4kOQSDrzEr+CENt10GUOI6K1ZBIug/2xuGhdv1IKosZE8
602LzigiJ0/SaD5fnzCELwcgcSLkNepQnso6ElXRc31se/TZERpFn0r5aw5qSW+4z1tysYdsaHiDFV4TtOovlO3UBvXuoR1XtpgY
O+JFA6muWwxWWkToeOZ1siiHcpjpf7UI7n8Q8j8jfzJGpSBhwF0934f7YXfyc2lW+vtehdHAfCjaQnooR9vtq0ma9ReSpveIj5LJ
42GWz/W+UrPkNGmrw04hgRW7RFTPKdngxFXpg4e6oj6kV9sApzsh26bjQO6Y5YOWskXVEULsYCH9XQjeEArJI/igFeTI+f+74Qqw
jhG9rmUtHGAJIDLf1DhLnh7lbJxUZ/ZEPKOzDa82f1uiaMnwpejUA5JYx/Jy0ftuJVmb1Pc3kBbc7EkfoJvKaWydxC5okaJRbll2
LHnrwofq7JA6A/UHbrpGKXrOXZdTRCICLGrdE1vjgpyjoO1crHEvi6jA11+UtNByNZ6wAFHyh5++0IPQud9jcxRDJluds5ZG0u2N
DMEtpBM4vWs7w9SuxCrrPknSeA1eqckPbkpoSHpn/840puuyU/ySd8hfyO6ReKaqYbwN67IcTJ5F1RZrRKykNs08g+RgErn0NyKD
vTpluuueMYmAy1CWCK98brabG2DuRX7/cGxy5g9NP0rdndOcNUN6GX2CeXHdgPhgBdyT6uG0HR+iZ5IRDzA9Exh6lB4RqTvG9TDm
4NtTCTjUsh5j8Y9XKbMHJZQ2rjYoetogjiXj0Njak9nrKpfpgDR+zt8rZIavpuLEJo5s5amjIlJaomg03K83ybf1imlIN7GFNttd
Fjw2jNRAx2YAe9IMbiM4ebhcFJcceWqR/EyrzsjUBymLe2ZHBnTOcPY2sMGO94bsv3YwKjxna6FitpYuxUlVkixsusJdRVzro/UP
Iph1NLsDbpTWr1chVGSkX1llvQPXYNSYrXttYs2mQw8D3Ev5dilzuFDcbUJ4pdVO87lkgj1pzd3pIZPb3NtyiUyHlhItdZC7qrs2
a79pL3p1usRYLOUHtYtqWbVY4W0c9vyuwyvR4CjM/bGxfV2g7clM6nYonbADAcPH9YRRbET1i+IURhCuHiz2t+cff1cE+BoBv50w
jgj3+AoOTav6prKvPzrVEA8IuOJs3hZuDhPr4N9VuPEDmPFRuA0cKdYFzhQqVL0ZdKf3O+5ig9PGaVw5RiZKE1DNvicrIYcgbh/y
jscV/9T1UmwQ48eJCSGiMN6tuP5l+WNAoqLXPzKNmB4QyBta2Cordxq3Faa2ws+xy+TRYfle2Y65jEEaXskh0O5Zfb8W9MfDkjCk
lAPbyvgg0j5m7nnNm8b105ItfDQhq6qo6sWKe7VsYGP+u7QVOnc7V4sWnVeB55gufDLXrI6Z15CR1Hl5IvFS74u/RnEBOLZUuZXH
1feLjrMbBkEkthj7ht9ynHoNXnyazFo71R7/b9FndDfjhbQbfHIBipzyJxDuCMvzJUrnxhU8Q0ckthFOAuUCdZd55GT7aSMJ1xzz
qjcyKVTQXuEsYGm5DgEO/zKJt84kFk/gxqcHQ1KhX6a51zYeHOYs+dDZ1yLvi9xII10MWJAdLiXRcg2vKbBWBzsnp/UJkUXWTE2Y
UOxXLzAmgUdomqDYweRfbitrKH5KKeOysgq9KudxMXRcm36BcEL1d4qARTlJksiLUGxbfsdTQYt5uBiq2YGJsrLQh9EniKNam+SI
MpqmKGPRkQQ/nLBK2BI7iY3G7Pfxfqte83raBlSmQSdaHIimRrZOzekSsqNAPoRHW8fa6/UeI1VBnF3qSRHL8QnoDt7CGndnxhnK
14S1+st8bWFzq0jHn8M4lW4yokbu784fW0xHcgJKB6XjyDKH5+yvY/1qfvRPldk0DKMR1pVk3t3v5DDnJA4q0pT66FP1HiAxcZ7P
IhnVkaEFNGpp7qvb6HadSwaTk5g+UhpvM5LuOVL3f+x4Q1t/xl7UHX9fRBG+rfa4kQm4iceF7Sk6OPQqebI/Ghga92iCjvK0cIXt
giZ2JtxlfGjxyfIAgF7qmOAaEeVxfhwWB+L/j3PTibGtjNZ8UpUBYXzKoa13A7NhvBBO9QRcdgL+hVkkypZT6a1qXz7Zj7GbNkDb
ygGQ8Ldhccp0baok/Gst+op3klTezSRfaiz8r1t+wVRoIo3mcVdjYOv6A8lRHPpKM70AWmykVfsCGN3Aqfvnqw2hO5NdcWYBNcSk
6Qd3Avj3fnexuNvkMbKCFkpxSjSytTG3BVb5+blpe0yDplLHkZZ4u4NmOnPoBCkhBvAOjo2pg93V8Jqlevo61Es8nFDsDQHjnuFj
HH/+KF7+N0QGwtJKO4AsPenqSgt0Vu1uFEpUKHw/crsQxnwC0m+SVl1CvofgNJYFd2hBqZOUdoeCxAdCJ8XSooJEIFxqyBV03/G2
EKJ07l+jlDRxs+Qu3tzvb/K83dcEhVCwKR44/XuwSLEbxtl6TLNOxJKnpFKWI1iUtzCuRqxsbGtekeHgYkJNd7XlPuG4Mlf0zVyM
5XzmrpNFBqlc6NWsUrCDai6dRMlwDnNAPNw1fmQkJpvYyoNozYcj7qbMyBgYqDhKK//SIZl57QPxqpczgsLHyekmPFq+IeWQKBhz
3GeGcFWcMyfjpioPCoV8UmxqLrRnfOfkoHqzsYQ/h5aoxVdOHnjdxd8nO4YOMFZLlZ9H6n+mTZZPWNsYzjQrfBRQnBGXuqkSWkJi
GqMTZqQDKIQKMwVi4OVOBLGz27aIHHTrqwVrjsvZFW+5EV5tx8ZBDpOMT3DIrB3l7KZ7nqQFsEjOS5cyW+cEkNKWebfWkVekk+ut
zwxZxITTSmZeTeqdLllE4SZTHcjsjXH+JnehuZj0Wiy+neVDir/zbXJbDQPnANE9uKEZEAbXmMClWSwnCvLxTJZZm/2WU4cCw0HL
liiMjVd1uW3rlTcstdkBSuABM3eI0+97cjOwEonDjFwJ69te3uj+3/LpBfOtu0ULwj6VE/z16KSKsTEuo/fJy6NJC6cCMf0t9Mo5
x3/RaHOInfzlN5gN697dJaUtFPKiFgmIbeuKyLc9BxMgMkLj4uTkzTyqla4MLek2fyvfBe0fMLm+5lhlNnEWcQvBu54dKOyXCIiu
oOwUmhb1wwyMQnvLgjzwzJP8CHwmY2DRsmbqZQjtu+WEbB4fPBtALgOWVz0Gi/F0unhZa9OVULq+SOy9q6dgWkU8P2idDlqFGnw4
/lU+lwWYaT4cfKouRL9C0hdom2F7CRf+NQQVMTsm68D23mxT34fFnx04VhJ/yQt3t5VIXgOrlSDUabfIMUHOhDrCN+Kj+p/LndNy
1A0Vx8aZw6hF/GQOY47gs4yKGHCyz/r8VHFqz7hZwJq4ooMm8ytt8ipMatlAZSJTRfQe61gPwXLqTjvP7iS+8FdayPBEPb1fbZJQ
Vm+axIJWFEJcK7QFeCBIfGC09syGuD0LtI16hhV2lVLGrCl/03WyqAomjQ+rp7dYnOWma8OnGAHpOG9v6UUspfAKnqTb7GuMFrmI
UcGgx+Y2ZfIIQUl9+B6W+VHO1r9RtcNXm065CaZyAcyy3ECAy1GmpXXiMiamHs0o1VXI2D4aKPehxOrQbaGfH2EKA2eT/dEfoS+N
q0tr+Q61ZwSVr+KW8M7eMD1eliWf2dyPLOBKYkUtnrNNEUQ21RuKEgJp7Z511en40Q9Bhjigx0w/4VQLxKglgt9QXhAWXFl+CEiw
TV0j+cE4pUJeDtcCoNe+trmnMWFtIvVgdRdGeft9wyr2X0FDsJ5+3xpKdSXyFM2MF45L9+urU/ibw4+re/ZPLtaHLv7gubKeuMSk
KILlU2bDlw3j8jOuuRHKPCiWPMuAJVJxv2e5zjxIRfOp+Q+p2amhtMd9ANpU3/c9qoLqLPAmwj0OTiUwp+gFn2i8GW6tkTkg2l1/
9jaZZeGZg3yuMK6nIH4QXbqgDffkHsxIyU82E4I90TnQPpuSvboY+a+10Fp7rbiMpbure7TV0BFmi66oCbS/cz8R8WelTU6Ifcxx
I4KbYrxoeJvAin+w1Fd+6gj1rIPHBN4BBJC6N1zzh0S5bHPJjGIllZ7p8B0Uxuds3GBuiLmnzvaVXCew5tJyJtuSzAj7dOT803Dj
hI3uuK6X+tCqA1AC482DWqVQO4jTSqXuVUdlh1SQd3J5yhsZjiisjL/9LldMV6pR9saLUm9SXXDCwuxMlwDhHAsPvhRFM4d3YoL3
XVDIvJlTvObl4iFFRPuJI1q6KlsJP3qactxsItb70co6quvAwTXK4taam+UHs6DUJI1Q8L7BG7RGOzZ5Bek4IpEZXphMw4b5qCNa
GqYXWVmaU4fFE7os0eQo4K4FPn3BBbAql12eHgCkOP0NWGJvjCNIslhrMdGnTNjyPzpEtIO3fwaghx4uAXdLEYbjbZJrcmw3Zj/P
rnSOSjSF3Tz9rpteacsByZpksahW6LhVpnzV6zmQrtOhr7iIWMhKS5Fp9zbghG/kSZ+o7yAoLTGU6segxV+0uM/F4UK2pB+wYg8E
Geos68UR8VzVMaOy/AR8xwITvx+lowvgBwW+igzql6cEd5qy7Q1fGSY/UmXASUt3k3+UGhUFqStjXla9dFmDtUtcrGQwUQPX/CwU
R5xlSPHr7xo7CjHwjfG2pagFmOlW9QO1X3Oohp0qP1Ml3HLE0DnmGcB7ngCKqtgpOzgKP+iPdDCHhmRMMjDw0qevI4UEtQxccyxF
VVJurzyQcNNGgp6PdZx+kdViKtnKu4dqi2y0aNXPevXten6hRifFnPBo0XXaNEcDaBJQE01FwEHnEfCovatbmbRmCEiAQZr/S8oV
Vp6NCFn/ybuAwjiwXCZjKFHqOPJEYVsYh0O6LKKzkw17qjYj3mk51UTQpqahH5ZDaymSRnnuT4Vw+r2H2fZHsmJ3H9pIX1ussPEp
CegzJWL2J3ZMdsM3Q7fubjjrRxoE7w8eeaSZn2oBi7uHUoXK0VwAcX6isYKQ4wyk3H7nsb/lekS32YneECKnl2HwvBgrslXf6+tT
MI6WaTqQAYqopfK5vx5fuZ80N67wMm1ueQNLo9+MU7sc25WOeR5d4UGfrlM/yslr/tX66xTE8rCT1WiBssJwb0kQ+LdCIyPEF0hw
cEqGv0D8k3Me/9LYwEB5szaApXDv0FANQY4x1c/My8egb/zzCt6JY/eOCG5ddiKTbAwRAE3JfcTLnfoBEqHyRK74LVo0a2ZjHPKK
iZHMX1ccHQpo8+wKsdu1+iZLVT0hDOaeAhsE1H9N3JmNPrnxOm0m3a+fhqAoO7lxTazn0KvZVRa7tbkzYdxJg6anOXdxArb1RteT
R8Up9LvNvxuW+5XnxFrDcMEEhip9XSwAvyYE/2/C2KtuS444f4uNPa6zJrpN54nNoerxGWQJxyaXqq/E52DBP1b2DuDOIz7UUdC3
SZUnV/8ugPVc0QrmDPfjdEsoU79j/2A8zMP2Swz5tV9Y2ZfnBjCRkvydFK/3YBgoNi2/Vy7ulVAESp1+6CxQotn1PzRWejfLE4dC
5kPNUVZbp2Djvl+k0jPGnO6OEk/mzCffkVYzGZu86e84c8erijllTh2MaOgYIyZNRAN0OeALRwjrw/H3eSNDohi+g4keob4prjRs
bmsnSon3PfdwFcMIOU6pwk4Ww3mvJcmj8tlPkLQKIwHrtAYIFq+wUswGcf3mZ4fk5Y61KVSyIV6T+ynp0m1Gc12oBarQU2SGD+wE
WpcaDGM3aTa2ypFa0QjM0sBY4B6EbYI+YKZfAOmY+xt3d15qio7OxKvazm2Fl8eIHPBoA9Yuw3XaBg4g8SDhcMrZy5jqXWHXo9lf
sZ3yNguO8vB+VSBW00CYNwf2JRkupDvhTgybBy159uotgA3SSlhB2cTqyMVRH6nSoxoHyODbfTTwpn5BCAhILECtpOYi3Iojw7sI
YNl0p/IcuzBJD1jk/gE20Eu8GChi9GrpPlmfHlL7OIYsNGT6ozsUN2TjsMDQDdP0Sm/LikZgNRVD5ZuLRpYW0V0mNmn2AXmyQMjI
4wLoaNbSp/SQ9MZPOHRUdEpftN/kMZdnlP+Bhj/8C6uZsWX2anamcrub9fhgPrFgOsHXzs/pUqKksMBwHyfOtqAs/V37bi76hWPt
r6QNZwSGLlf+82TuvxyqUNgPisc2kdg4eTetm7UGevZCEQBT2Qhfg6WULricFwUEUSygvKxZZHXntKPPy/+/q/SKH+RB2PhJeWpq
RavNJBjIPp2AHTKKFchShQyK2EBY+WbVdy6DeUsbx8208n+1bpWpUfDOIyS83gBw15t2VN7UPhrKLkzNtWvgilT+Ld81KF2vP8Mc
ro07v6KA8sEIJFjiQALkB83dXaLYSCQLYZ12sUfIO7/h7UNP2+8eNQVVE7+1chK0HfoHJLtH4/juuW9wb6wtVdmOJyjyKqAQAyeq
+4dqt2/q6LRLca0qFlGqzUPH6TkQhkJVEat1kKi6bRUmDB4uslY62XNIQccnzAIjrDi6mj89dFgDhDiqzJj0baWe73+yW4ODoBc/
b+mao0hKMkmh5jRu5tQDdjBTnbuS5S5xivxkjcXm0bNCGbPf/GNDXQIp8SuCO6SvWttyFsWH5SxxxaAIoVbk+QbpkHR5yA0EpTGm
toD5trEGQcr2WzKHgxSkUIs6RUvMmLHLInuDSum1i9BLxcDOziz60bSjfgFexHICQt3ogIgy44hg0YLW0Bp2EVpr1oQlxgfbcDBi
EkSXXcu4C2qlX9R7gTC2AH/OhUANY5thyJpNjDczziWYpP/8Oa9sDQhT6Ws9YNhc6Us1WgGIs4VIMziHH3U7bMNUVjrxyTfSJo63
wYdhuwnfIHNv+0gfTGKlpbESkN7v6eeUgs2k2H3jS0J0/nfyZ2w42AEcwbLfO6A9kviJTBh8glaSyxiTcWWjWkQ1rbDg5yiFeUvI
fgUuV2iz10ab34JTH4bJWRTHWF9xaW4Y6i78xrYMspKhzUqozsrOhJhURtqoHT0BflgMFW5qOZ0ZO/7ILEi0I5P0R+BUodJJrb/x
xqouQRmkT3pdS0DNf4q2CGjJOcz2nE5sv8QRkVjZB1mOteL7ene6dQT1+Z2mZlLpXZDS8jYwDAqukcRkJOnM8V6cOu1xqX6goJZU
RN8O1LmcMnXN/X0ObYD5r/hCwy6HlZagzINSvzQdl9S9JDQWzOUceelI/jYwW88xtfzXHVrZMMj5IzkVgs0gWt0q7YhSC3iJQqGU
VVM9dtiCv2nJj174SQCOBZ3CfXMBRELBKkwKrOoGIx+DA0Ps7EIoeeiNp0sZevDqRyhC2k4ppUJrg6DsydnLOd3TVB3cwMKhnx8Q
MCCKHhkRIAt7ZRBOH3NI8F9fzhmylZ+xhgsTir2ywrfpDV8P1/HI9x60WWdqq8bWRF/ioVixV8wy2NT7UlhEDXzh78cEpBbC1uoY
zV7g2s6sQ0TQeZI6+P15O94emBHVICrLiW0TxkM3pLr0U6W8BKTEWw9GThzlkTv7FoiJw9nFGA5/I0vfaUtHOazB49YVJQa477EX
KrStNmWvTF6NrHDlyIBgE1J8Ul8qrvC3itd4ERr6EgIFzdCDxBAuCxz2qSNPVhuSsgQuxxj+hVoVsW8jnmu9zjkYhtxqsvMeWY7f
EN4jmsxFAUAZT8u/lPfCkBKtzSLPaIMvARkxrdgcaV4v/k7NtYODDnvHxcOo86vdM50V5pgymUP+SbK+y9bA7pBG8pzXkGGBvttw
4Skwhmtn4t5atDcLqRlFg/p2JAFxiKR5jk6rWZzS8zW5dX4jho+s6zlep6DSu4pz2RpPTWMs1PTrLSMsg8ihblfr5CY3nQTTt0Fz
ZVQB5A173nDXRmE2lewAu8639Hrttug0BuM+iK7Vk4DzDKCLE0kf/EDtWF1VGx6yjh6eVe1BPAJu8cJF3PDbohW3DJDsur6Pyl04
YE/JsDT2IotsQXE/3DAhSCbx6qLqopayUR0aDyjXQcwg0KKuOHY26ZlsBa1vm6b3bdkcfVhKZvz3bihJOymjoPy8d4R8HIvF28ai
9vMatD8y8ubzfj3FIafufDDZsBOsoK5ZxuBRW9w++9YDrg7XsyUnzj2OL3KDOlxWXL2bopmaKcu7VKW65FITNqq5362EV/UCUsJJ
BfxM/j5SvZkAzUbxftBLehkGid1F8t6KGWFn6W476KHz/R4/HfM87gZ8zeZAgE4TGlizcEqNGaHb4Fp6B3RFADbvbO3hXNzZamnX
TX1wl66ik8kUdv+FEQiKtofnAi1qPD0XbR5D7SvD1a9p8iP+oL3L5HnDC0Lub+po2hhSc2H7CwpCUlRRS3J2sGGVO5xX4iK16EXO
eOLupByhBxP5IzoFYfBsZTUNTB+rN0cmn+SyYtcnvwvpg8twctAuHhGK5Hi7vlVGGsr9vcJmP8haDeDpJi/PVum8vQ3EZfnhXT4f
BdceIbRmwVBIx4nDZE3T2/8HWbPB8JKvpPwqb4CwkChhoUL7/HAMxXvqDB24kJbu3I5cewB97ZfDo6IkrThCFMdkCk1XdsuIOkji
zUvrmYBIFDIu0xs/gxXR5EJ9HDIuh8eGlXT+XRgGTDOkCmRSOuKW6SE2CmE6P2Gi/GJcE44pzYw70sgeVrkk5k1Ni6kP5MHrCFQG
FhsUCukKXOqN4q0fGTabmPxdFrdPM1jaxk5dhgRDSC7+/kjZPJQHA+B0s4+FE9i1FdfrblCzlNAinlzN9x9mDiYs2eovhLK+bpe+
ElQW9nB8W6BXsYCqo94w/bF1OPfTCjKTCkzyVDPCu1t4HKT6okfoD13Gq2jjpD5dYeFVW5NifjtDLVa2XxWA3A2aNbP9GltqcHrt
v7eJzP5XMu1G/l61J7GlDjXm8JR8qAWn7dntl/CQIWn9ZJjGzFUpC4xHUnMwvnzEfcQ3CAKhdqFV7oO6DUVFrSxHgGjaJvVESgrV
xDDC1OENJZ5QoWCJmCRbbZo5y4T9UYh3FCP1Pcy2KllFsjTbDcVcmI7NmLVmHNDH1ei0i5ukH6ykqyKEmMYF2Q4kfC58qECycGyi
FRz4eRIWRbE5h39K7gPTnoLJR1l5msmgHPbI5FXOg44ADVGLfrbDK2/XD7JPMpSeyrSDqIB/Hy/2wSDkiAakmQIVPzhzSKhTFTmP
5AkIQQgjZaSe4bLifw7c5tWRmgR/ViaVmMlemQh9MrRJWObhq4K78nTVn3Ar33iTg8JceuIEULEbYQiD/nvEvts0BdE4mZ1ZM6lg
jEFiKe4mYT1znEYBCo/9uyClU0cROtBX1jZ9aCI1za1nCwr1xc35mLuzTqV7CSIDM4W+AK74wItNPJyTtCqKpznaCU3E5W/bkKMF
qPPT818p4fPn9Wr7Af1qD49DlE1WtigDHJBB1a4Rfb8TUjPdqhfWww61FIDEYhTRYQmMHGM6HCySw0YEGaOFF0W2bDl56NTST5Lu
4uxsI4/O1cCcHstiw3VSUxjytMGPBUs7HuhTdVyrPEDdDOUPznzHjnTYNmLaZBgZAZeQEgULmd71tf1S2aubxXsCK++FDBcKhv4u
1WR4FC4I2wBzf9Un/Ad1VNb/FIcGl30l5OBWcXtp487oOIXeo/7Sw0LPQ+kj0kVjFGZ2P8CvLTsuCI8qzdTpfnLlZrQBiTO6tnGp
suh1g0Aev+ZQfwvciXgMR47n28mhswdAZq8MoQTXxWsSKpGGAojrhZZgZ4DR2Hz1MZm+hAI5j4N8mecp2Aj75Nro99GwMkkBFtFk
ARqtSqWD7qHWjapwn6EiHq4i90lmw2ho8r467ihIbrcRcsWS/lK1wyDz2kD6qgxuq+jntIEeitTYJNUAXJMY0xRy8A2mO/bgmCfq
P5wiXwhowvq0fvU2PnYKGPzmnA5nD4tfVdqqKUWzK2OWh3VsVtr1ZonPhEVVVoz+TYwvmFvrH78V7CfDd7CnZEZxtL4XH/6QBoCU
3XuXnFIuhFyMne6Qz+dvFPqa7U9/7kaeSO66qE1d7yk7Ygsn8i2RfsMunlt4vDpe9b+UqlY44XRKWzNTZrO86nP4xjnYBIf+APiG
ToDGsegohrJp6lR1OkzNg2e+IoHg4UttxaZVQqcXyydZqxApahqUmuABfrZETJLyiOlU1MVQ8FP/zGi6A9cBB7oB584MvgXY8LF9
IFithaQh+N/G5l1tj2PakjAyTIu3Q4HRo5zgwtX1u+MZQIsSPYFfAmaDuz/RjaoYOCY9LF+qKY0KVeiggS1roMmaD3GWtV3sl9GH
Dy520Yz6aa7vWAZZaY1AzLZfRWiOT6sz78ammThkLXa8IqJdCVgzCq2OEOk6mJT5VrAvUpCemBDWVm7WPdRLxt7hj11A5ZyTIWDq
fvRKf6wFFfSO5ukPxUJqOsSzNYIXsUK7ZFOZyxV/ma+/8PWEK7/68N0hH1xUTaEcgsOLlJ6apPaC2Q1vhlX0E4Pa1FbEhxq0z40q
4jlMuuwOIHsTs7JMtDIig4qvMIETuQHSceqI/3jcEBEcanEYEX9vgfz1IoP60Mvioou1iweJqqGg3JptU7xXEvtYlr3rPx2o2mqi
KXn2O2Yty+dFOPI8P7/4V/J6Qg5DryAHAzlzmOk3MDsYCm1VD5Ua8FQ/EO2QhHS8yobVZWz3xcihhVtALdJBt33JGL+TMPC9K6H8
tNuu4TnLCu09Zoff8KVyE663MjsAob6ZgwDa0R+mIZEfmP1if0oZLulG+VN+KUMiwyFPXePqC/Mo3aB3cZ/FH7XaVVZ4S4VKPue9
dvJAVyT9rlGhNQ9/Aa1VnJ9TE3wWNPpL1Gv+rF6nvnMxeIA6yQs8sEHpXwAA5Ksgo7UMe3mhWpLjylkH7NgpS//Pp6xc4UAysbTm
1IcOQE057dazse41NejDDWmKxtUjCIs+Ijt7jOO1cJYCFoWIYyxO6ubmmIWEQ2RCL0q3bJS0xQUBu63rG2em9Mz/cHeoug8gumld
x3C60lTuLHI2g4lOU4zu/WlMaCLIgYjMm69eyCOVMTZsp9jJ5MN0YQodOSQHuU2GQru5PFT/QmKJkeAieDP24aleZZB6fSnl7tr3
ICV20Kbu5foyVduJXIERpbplcpHJc8fl6sGkJQK5bDlQwQRrAN4Fif+wauMCmOGaEX9C1yvZCqALQy7p1pDBu6O611YQwrQNq4r4
ZH51ppFZjiOrO+lOgcWmTHUEkzHP+GTEdAKGXTtJz5UdeAH5VfZ5oNgBa7vI2xG8Lz3EZzoi22oFrl+pXkwCM6U6Gv/WkBMaTIut
wjTukCY4wyzTIm1VVoElAkgVDvScd0fxbSWUoGKPzaqqcovv/19dt0+MT0KcMbEM+IUviNvWUx1C6N5BjpqXGaIiyeHlh1/0I7f8
s4bM4fccJhr2Lbk2maOhyffbIzlx4dzSg6940eHwA/CSTHSjTEwf74iaTLvWZcfirxsjpFwm05VxbBhcE6OIBKOMQl8OA4hWggHq
j/L3aEudpw4TyDo8yQSwGa/ekBVqFUn44o+4Em6mm1AmMQ2krlgQeq6XquttOXPnljFqBcS8cQSXAaOidNOhWOp2yDd7/FRteLlO
+i7uw1QLaF0/AlziOk2OzO7z2/M/HngYG44aMefqVDNQda+hmtkedsO03ZC+Sy3aHGPBRwiTcJxRmareGQdQquplScPlWlfy5eg2
87ONLu0i4kE+OrCEmrnPAh/L2b2yA6zNpzqM7qZLAceJlCz/B4V/qk0dABPH+JtHbnQlwF3kRB0d6DVZCFZHPcD/JlYf2hkhKjcV
MjFOo7YqCqgcAxQBMariSBAsx+G5RwfBj7XKkVoc6i72nHoM7dCaafUyRxsK8IgXmXlPR2XIIkpZiapymKiX5YThcRRzDfhOosz9
liCZJ73MDN985L7/eQvDAvDVn8Vefwpx2mB8QMQRc7DQwdhLcV5bDey+kNsEbjQo2ohT5f4RhGr3w4oEJJ140s8eM6Ouis9Ox/B5
fbW/aVxOFFaBVz1PJbX7vui4d+vmZRkPyyK1QVtyOM7S0PZiIFZUNoNvZ+QlGW9RF6JUgHkeDWurlVx07DgFVlVHt0/ziI9NsP+t
HV7EJO+Ur3TL02DJtnedR0P2eeQh/fpjasE89Lgxr+Ifu84T3SVkhPTVF4SSE0un/n/yT00IXFFr4NeJd5kOnvrrfAqHPwybtMlo
SpRTz7V9YlLUsJqD5xJQl+dYSuvlJF+yeAx03E0QrHwNz9xPd+xJKdU8ZTRKoqzKwWORL8AMMpTOhk25zSrQEuV+6dM37AlKhqyF
EwTVXAplUPglzRGYyEuNIXRBAB5C3oZC1Ot+ZpGAcyCfH4t7jgmEz1XugXqOTiKjOFaYwOqBhtny2B0wXQFxobHZI4QG+xsVHY/+
b5wCOIEX3spype3GOQ3SexcGbZnKIfKt+QgTrFuQrPVe4+rxRch8eKO/6apasaOQjsjBgbkyvOZNLrCZCsHt90RKQeOGmRjKflR0
gWZItSAO/ZhclWeexuqIc9peEtJphmDm/a8W8rSfqdTWVqhv+EKKlnEYXm67mcFqwywVStyAJoXWIkJtM+OmYBC0p54MCmOkUp4j
YKmddvbAEnBrZtCmB4K6DdtaMtKEL515a+w0cpDifttEV7aLrQ429ujvALvlcHlGtIadOOm6FL/pLsuKsxpX+r1TNLhx/+iiElag
s12wKx8+/p5kRdLTQYivarZN7rTZDzU4LPk95HJSHgywAARHZFD0hNWEwBzlp447cZnifpSrTSV3DjvG2Y1bo/UReGAs5Kc0E+f2
xgOr2Ry5vKpJ8mtoMd0ANuGWvRfcwG92tiVkeiitOHyGq2VxnGYEUO1APmtL2+2xV9PslNes2abZ1LNiq08ZvTMO+NF0lqiipPzw
XUTW8l/PKNVdlcQ9rJ4gKOhkVjht4Cm8305bte7jwaKRqiK8Do19mdLUw+XEiCvVwggMGxSC5/uBOFf8QloPJwWrtsjhUQ4yFZGW
3rRMhMLI4eEAg9W0ObHgDTuRJwTTPk2jZUio7ZD+1JM8FaDJZxhMyf0ERopsXX+8Bq6fUMznGIoQBMZRh4PTr3KaQYLjYsvMWBto
X+YcP8GHbAn6aI6SIfBizEOSXSJdISyiBpLPHwjS7YLI3E0N1YumX4i/CASx7d+M+07FOO4rfUdJXpgP7EX4/ymG6cKyO4Xbvx1V
RkuCRzsn1qf0SJtBpKa/v9SdzkZpQ7UNoDBR4sZ/5Hrw3fMnWcHhb1hVXSTphtCqDCX0Dimw75Bjae1DNmYEVABVBJrPcqvbtpUJ
1Pu1RDUQzDlxIL76i8kIUob7dbQFOWW92rpPI8skNVGF9ULZlKzK/eKxEd0IyFnDWXbc65xfdWwfbfdlD/1Gbqd/NHFQzilRXfiG
buY9UGK6KwHp4yeNuGvnQcy/vvpOF2dCiLHfg/D++7AnSw5JPAa09b9jQK8BXbIGN3awOIZvXX+GN9Sq7D8P+nbZe9kgvKhFtOkl
ZAQ7uaAT5Na/obTRKctHxe5Uo3jEQi6H6yfZ+61k2azlbUAnID0NvXNWrqqxkx6neSBTU74SywqX4XXPqflbJC68OQSX/QNcqjaD
zryOgU0BDZxn1snH8LlZb/uMVv6DJAZOaJ/jjNdKgcYA35WSOMFkMsPXcYJ/NMe2uA6xeX6p9OB2tOoeU+JGZU3IKOQs8sBcYH9Q
tgKzibsb0hp4+cKkFGCBFaC+oKm5DGyVsdcj0ZhRj0qu3twtGIzPHy+Hx4fSrivcv/Zt1tMO3Z+3ZWvn7Rm066tGka+hDiZcw3ut
w1LA818oe6qqMOjmnmpCvqE5Uf5b1ZPz/GELstnjfbKfID6so0/sVX0uePrepYdG9AOPHEEl7oTKnSIPYLui3fHVp3m2mW4Kf5Xu
iBW/oi6bj5BzruLn1mY2uFZ1ku8JLIU29Jkq61HcDch60o5zTElhgXeZwP0P3YBUGNDv94FVncoiVwAgwvXoLnWLHwIsn3ctVwdu
SAdktQrY6RH204hEKRGT+0nF/5XOk67OoTUnoimxTG+ZSxPZcQUD6bGgFdBtzG9wt/Zi4hPI8WJDVFpF3kCd6QiiDOtrmP1X67Bd
X5NMTVwYHABRUaCel3QP71mS257WznqnTt/FaNjEm4GxgXitIs/d20o8ZyEfyui7RpeqmmRLi2fA5IbIJaJmROltfiW5uBXuQlV/
xQ9gELyX3TEkY3a5Hct4Engf0zRmnceb2CIwoJiX5UDnLWmZ9dyRLQ3nrB7udBFNpNdMu0ZX/LaAdft82Z2hcyYWnYQ0tZ8v0syE
NnFx9Qz+811ZzH3d4oZnfQj3QYY9HhJANr9X3NoU6vxypLgWqHBsVfADBKx5wGYM7ZZU8pOhGVwE2cC4ozC6DJEGRxyMCzeWWZ2s
46J0IkJfqaZ+eJg3fY7e6rraleYd1hxXMS4czW9AyIeUruHJeyqb4MsqWcm1My/2g92Gx124tPulM6drAXeRqzLf3jdmBIFwgrOM
UIW7m4NzyV6RFRPJ3MnWLLU5pyVvUzaqkm7R8AZ2QNFBhOCnthoClotKjJAE2uSHpNNloSkRQaHt7Nus9uXGlUyyftNymmrVk+qx
aoq9qNpDDdSU4H9dHJo+HzSlzj0Zxi3vrFEmZlItOmAxQK0EAmFYgq5qsWozNp82RES+oE4Werx8ZL4hkxyXCzQbixjGxV0ge5bj
uAw08Kn6rbLcGTWyTnCja5fv8eGPI+bUFJNDH45qXWbeCW64joyL+cGKhjpDOZwosbm3/E5D6Ij5U3a5bwv9N0HRCYVLg0lSb8bF
8MdH4pfIhEE69yaepAAKncDY40p7boj2IHCyUzseVjUuKtdBGngjmY/wpTzBWejfbNW2i5EEj+sUsAOtBpakdH4MJaoS51bX521S
T1Y+IPOKF5vvVMC5T0mhyRxoVoGWFMz/oQkf3VdEPoX1nLg0trxrdrBbhT4lSAVYoQGQEEkS7ZX0u2DmnXm3sxMb/km2YSh1ULqB
hJzynJj3JfKS76adj3y2FtRWvzTZreLtzKrcITTkHFolrx5lpVZsN1yArLjJTIwBGe+x2q58hTH8W5Q8K54g2Re+DRWbrDtxCeKO
w/6BCKz7PASV2jRwUgP+sDJuaM6j/q5nULITpbrRJe3+UBG4/PLhQk/8MhtNZhGxoKgs5BRxbzqJqtW/mxTZ2B4dMcZf8D3WWWI2
k6GhLzYqNYSnXV3YrHUY9c+ImMY/b32KTYbUrz5MJNHw+EaO/pLJzX518tpGsA3q/+PCQj4Osm+G1aLUrfvHu1r/hXs+0HBRa3vS
2Vp87MTSKrDXcHRC/80lFgAU4ctsT+f3TgEamgtT9AoEcYV/HEnlFRcTFsmK9ZMjDLtoFjL/a55TOh8099GEHiezQn4FoV00gDB+
BYVmxO31Br23vBgWlvWHo/oNi+jaw30taeyL0+XF8wtbizuRrVC7Awi1NKoyKHHeJajEOw9SesdAqRNNkNjOJwJghwaoDN0+GeDL
6PQgdmVyEnw/CTN1S+TQqNs1TgqFQ4tKxWQuxDteC4jeIWX04DAN1Y0WuW/PMOwjmTBGz//xTwwQimvZOzXZzx475D4FECObwKfb
knHWZWbz4l2ftPR95IkdX1E1nbAjl/Wq6qjIVr4jruOI0r6FERYq/gvFEzOmF6JyMV02KMYRsgMAxXcgsjVaTVM7+4s8QL5sauWy
3pTqZpwV+sihHAf5FYr4Dfa13yngdwL7jTdRzor5ujZH4uzgf1P1/vzm+L3/DteMxYonI5T+YmBGe808vxO2BhChYZhKyrW558mb
3evFVYZG7rzOuCcjvdJb63dW8omo1KTXTJEuVhUC+DnRSjVIdi7fu4ratNQgCK9PkVE2eTKl/PjSUJil9bk7VaB9RrGv3j9/suZJ
8j0/w+4569FbiqS4Hf3+rrK97hbZyM0ecewfDI/pSG+4XnvJ79/gMxCrlgm088pauSV44ZrO5lb9e9gBguxTZznqzwKyaDvAqBFx
N6ZzshU4rpRuSmhGkJkTCxGXH010hAoH6hLhdU0Dvw4G2Uxioes3LNHx5tojMSmrhq8cBnjjcQjwT4cA86RRWXGeulh26k5djqMi
OU6fslS3FgD+TTpcUh1Kxo3URDq/WW6hbnqjK+XZRePAygxhvE0lGoQ2UFk/ex1Qwjyb+xNGGm1SIZbnMqQdSDjGw9WhxHN7lNmB
Gqnmo+1MjhJ8VxVIZ9z8VmwAfsIJQo9hCY3sWmN+F+1H7KgQi9RC+jk+62eVuQ03/cC0WMdirmqmeLo3Xxs4DA5UlGczwFHI5Xnd
3xE6APM5v+kAjBjhY7I+z/XJMFPn/GEG4XQmhbkaeCi/cHRDUH0J1dZ2sxA8VeiEIsXLKfaj45rww9GFILw5isuIWemILbDWXgWf
6J0lQm7S8UTaUgzH1G+0uE2EBr/OAZZnlymYQQ1EFUvER2WHiB4jgc2akCEbWAIANNedJrOoBKWEjkcjkd6tV2v3A3p02WE5OB2T
QJg1KJnpROEFY1cOdCGbXTOTLJ+CI/wfQ1ADy340tEBsltLEGGBhZNtDexSna0QcuBFA4R12y7tg4IdHsh+k0eJkKzCfAU5GpY+U
JTyqrb8NJiBDA6q86DdVrYtRWGGy5n0JT1H2JZfX/aUc8pZANqs79TE8DdG7jQO9LNy52belcnasQnCLqG2S82KdKbHtq3YO7R6+
G5Radr2poyvzqOQDppQBFxqrFpq49Rdzazc45AjKvBbWxlHjtdaoP+7bL4V2wNJKnaiDd5w5mKvBtaTdHHhqr+3fwd9mTzghHdlg
PXL3Pn1H/d/rnb6BPPdiRv3c0nSxGFwOpFM3VCgTZkh/2Rm059jmL8wmFzD3mHzKVsy1B03BixaTgttXLrIjPv6gTkZqFL5Hzb2W
O+x6gl8c5S2i1ccXAkHRE+Fw0dciucJFXGch8tMBrsuX32S4nuDWeLtMraTAFGR/WVaLuRqaBOiOORee+Arud6lq+polX0MdA29f
gWwfFmA3YCCn/20c73A+CFIxz93i0HYxSnHtGBv3OY1774C8zJrYkE01sY6LbfavrTgVXiYCfVca5u7t/MSXNNnQYq1XjQQq/GjL
dKupjyWTT/xQW9gSCYNOkyWls+Oz4L+6JpE4fuQrExbX2/uxDi1Za6g336r4R66ifqycf/DV4NiiGB61JNr13f6Qfrq0Fzy5z9hD
UbT1YToiiYIv8ADinY9+Ax0PSFDvNDepCnakNpZtzY5mMJ9KajZB5UH1era+QpiZmwYY/ZYlOKm+ROWpOl9t05MHWqwg3j0tN4y3
8dbrqgi9vpEDVoVZYT6MY3Ir0gbi5jNxxkkIvWZT/IPC6RsNlA3VNmHR9w2zUA4D6APmlYx8Lob8WbpsLeYd4yIwW1ttBA3eynuC
QOQqjnUbMIlI6ibnNHgdlMvt2DRsR/F8s6gJn+/9pF70n5w6N5pxOeNWKmA9qyNrsCBruO81hJsGI9LCl4FFW1+wAM104YgAfjQU
UjK683y1Mvq2N2ve5uxML3RsXnu8Y4jF9RqEFzw66ir9UYasltx4sZOrXi1+8opka7rtAqnims6kiGA6Oab0xbTZ16aPgYdRYZ0t
gFyZO+9ppSnkFz+ecdcSEVld8rkGHlYzAzj9DA48BH8dzHXXNkvvGXGAoQwI7o8FktW08zJqg+8rjin+/Z0ugqq1sPCyYZOf3w+j
pJpwJl8lpr4fpTCDkfA6Hnguf4z5/FgCwssbHvSMa7EHtGtcq3/XC2wnslcdwIbnhrma1unkpdJ37JoW4EIWbOVUicifHKkDDjPC
Cc8ZOEfebjJL+3K8v1zOCYrsKXCUUdtBTGORWjB8YoU/a+gG0/Ptv30oJEZ/eFRE3pOt1yDCSQcdKQ3nehiapYcpxmySFpQupg60
ijUzNOBPVNJXirEGiWeHqhgfLf7bN1/KEEUfs4sSsWMnqgAZPss/eD2gPEJemfPQqtwPBtSIT1JCdppptyEOX6L4p5ZZGTzGP7QB
QcqfP+2GnlTwkNVruQLwY2UHuoM9WYclzqHq2DwXfHrrTMAFD0pVYcye/lrwiSXLy9Y33N7yhJW8TYCLB9KSe/4MblpC+0kbp/IV
r/XN0ZWa/NyIunapdaKQOKmiQO8lPoeU7oFp3PxZVnigjCY27zQ0+kGxrMEmnHmVS8TM7OfgYAV9otZqxYrE84mYYX2CryDdxlXC
vwBx5z8uIIVo3qYXYnkcg3ETy7TQCLuORz2QwADwJ31JahvQlB07zNqVz3ZU0nNcBfI4LS3azTyEQbQI4cvA9TwDa39W9CHbjEw/
4iefMLoTkVDbe8GHqn9ONQ17aIvT8AU4haP9TWvcfcHImPztDLKUyUvsytx8uUnNJ8mZAkwK90kxcZTTipmpkEH/WffplCyZ8qtH
k6t1OtEM3FXZHaPLdU+C/bZgn3OUd3fNUOt+sXZNUGfOfN/JRZgjmjcIzk8rQeD99xBjish3PZ3DtoJC/oCRK+hy345N0dnnJExu
RZhRXZB96HHbA80uU10ItfaqnvfjA2Yhy564w8Iafl5qwoxsvlh2Dowqkg4d6m4K9ga9/8dwhup2lMDp2DMG7C2BHkEuk7xsNKM/
KGCjitlHbmOqrumf2IS+AgaM0ssDyfczeGN2wslCIox8vTZnZt4UOsMcQLkjb0or5+VTpzHU62e7tfqKJydj58D6RmrnncV/Lfzy
Wsbzxmw/8dAk/HxAYzedWY9DdsZGe//0p45kjXdl0OSVi/9P+6tQ578tjuey1Cqy8oDQuxkrVC9fEL8h6xSedlfIgvQRdRnzFNAA
PLffsbpXt7SCp4vSy8oXMEl4HtHd9rE1Wq9WAB9ETxW0nBvDpuiu/OesaG2gsyq8TKHsd4NDk73HoZwSyHDLAgNwh/rUQf/t872V
sgRKUwCoEYKLcBoWopWTeHGcq97kqjdPeF9xPrOnu27Zc7y2VoOpg/e1hLr1KBN/9ZRensL6FWqJSINkuWlXrIaIVdBsOQ0wpCju
w4zRhl0FvyWbTc8EVVQZA4VlK+xw58Njdeg1z6dvyooaQSMLM7L4SlSFBdNBH73nCPe7ezYRAu+ny81oVfhb+v0TL/5Fi0HZSMjN
gvkF0oFztKubvfpS4DkWG+ngB/8/z1v5MFahvcvNPzJ+iPiAXI5cmcqhmj5apWlhKTuu2rlHTYe19CjU/v8RDU1u4PWNDjGNfHaZ
RfodM5bREy8gMRcNQi0zhawuR6zlzVwlmCNHnvPZDlOTFmGig/GXYGAnfsvmJ1qNwXhxfji/XYlFfF7bL/Q4mMpRWLbl818Ro+x0
z8Naa4T0BAm9USbWgbK7j2Dz97tVuLbDYe4cpfSIut/rYEq1Nlorg1ToM1TZBuSvHbtVozGwKU7AeZdFIxHay7njfFQxFs+gRqKy
LKRtQWQnN7eBEWD4uDtedTmcWB+5F3R4uamnn2agHx1KV/9AK90Bu9P64YDNGNq0yjSh6op0QTMLXLdJ/KiFg4WipOBMTcyDrhmF
PhVuuvzmaMoTOV1rQxjUm1nweqJIA386Vj2IQZHRiWmAmQt11RGWg/8ep5biFZvkab1QLMY1ErKjUJ46W+uBmpRZ3J5KJllMWxrL
JkPiWQxGzvtsQaq9Xl99A3sZS8nJZgvydA0NVnM4r2Cx/69lFsymBXBuCaGIP3HTHRE5Z2avxzYP/xOvGCmDx3wrVKicnpEhXYt4
ujNVBztFDWLcP+A1ao0uZACUCHAYMp8KXVjrXGepPPC099F+Sf8wR6S1Ln7zoAtBHyd71lHfgTdCgkhITaA36ft4y5SHmZLgc3NF
BeOC4tVjUYSVFpZcRiA0WMJ7zhUc+4wqskOSetiPvbQ2j/oUYrzisHqcMxW+knidO0btr+W/1lZJS1Pcf6ABqQLsx2H6bDugvnAK
S7d7rD4CpkpdmlsuBDsRGX9kd8Aomoq6nCgASrVeCzs8Whh2kHOns0NZwxqZsBEZ80u2BEf0gvc9QEgaLoGeXSY7BDiAA6frYrqX
kZGClQh5NSUkgM8nB+4X0VQ9SL++0Dv2E9i1Qr9i+yLuMcJkuaeNvsoXaJz3IC8THnHz/5C8v8kUn+Ux+sOPhdhg0X9H5gIbWcNV
hpARaOOZ1xaoWeqbFbN7YGY7F52+8TRUnk24TSl6OuSlqKRjZGNr4KnXqP6Q/ekwskC9yrTmJJMZh9zs9jz0igvY7V46KOYKsTGh
v3wMn/kTqudYgM1D9tZC1jG2QI0pc9kkF04A17AbZZQ0zsOd/bnZ/TOgLQNGDGg7HrkXld3Pjyc9TNEVGwpZsqy7/GzLhQOvTlWE
e4mPmjEjcq6qzstDb21zDGH3h0czFwQNl6eoLAuu8q3ybspz4s4BjIRQXSFG4VCH5/oWaOE7X0y1WLLFYd0/RlhPXjisn4EFcJgR
rsbtT/rrpn1IOkBJqQq1lmAg/p32iKW3n/c8v5ihY8ve802Tjy/jKO1LBtJu0M9jz8ZrI3vwDPr2M8wYzgYkR5Vy7a51HLdg6moA
xeVLB/NwJNmrhFfYpPQ5tisSbJbjLx+jm3v5lD76MouPHdgFYy37TQ2ID6aBZkzRMH2wUc1LLVuAULjKb1OfvNbjRpnTnfsN/W8W
Uk2rrN2jBUgL+H2VZuD83DqCjSUQKefdfwOk3/vTNZLip2BZv60zKIis7abMtpp/1V/lW8KuteRxett7sRMz0cV/VoOVTL48tZEA
4DkAHHmQ+mZAENWCZzQYBZGoUXjn2nnLyW0jL8EaEN12eDuMszYs9gth0V3j5DQl+GGM1FspMiOqECq/VZaV4tg2PYfeO2YryaCO
ydCIjUZ7EKfmdJlJJh15duOCj+ML/LggIYoEeMN9WhtsAGIp4joXM+2IkaOj4Tc/8RCiTcVd0hbCCZ22ZoxovBA+tAmUr71wOY9R
ncJ7zhQdMtxZLt8VT3jyL9WtWes0EGg39BVUUsKv0RYgiYnZ2vHLdjYh4Thb4LeNN1n1VyMeKJ+Wz+CitvzOW3s9HMCLUzERvLcp
jHwmk4lvZOcrYk6d+kAfRfruzb74JuZvUaZrZtMBqb26xCrCkQP5HF+Slz6yLmKCOC/SxPRZaohy7jyFGL7qpbzSOqlhWQ/I1N44
J0YUx2fxkylOcAmngl97kcdnYJm4zRWSkEuYBvq8TzvefyvqVVYQhCMQSnWCjWspPLvDSYJylngoWjyOBIGfSDI0rnmIYS+YP690
TFTZvJBu+d7O/e7jP51EBq7D3TQrjXFuNw9y5JFPrYjJc/O/sIrQg5kA9JLJoDf7ZXMWLWlZ0rzFPAc4QOYyq/sqYzSI4VT40I66
qUaQN5IMcJkx45eDBk3yH4AT7yW0Wodktlxyr4qmyBMVBCgjS4DsQ6W8Va4GvOjP34/Gv4nv/WerbU8ps3VyaFr9uCwCTOEZYPA0
KInxSZ3mKXWUCwBQEASEL8vhqbdv76+9Lc29v66KIUG678N+yFbb8NTQn/jF37CK4102gyMM/BuIgPTvpTKbvoFwtKwxlwj0t7rN
/5D2+UuxOEcrx+mMopoaFfNu8zO6DbzJKIwdXxrPBdUgMt9eo0lPhA51LYmLHLidrnm2OS5bv4fsoEUwEPfVkUSJKqgYiRltJwwn
HT9Ki17kwpgXmW2IV/+x6ICrcb9XQhjD4ublc/9OeHQXShvuCcvlaOm6QHHg5BtdoRr22feocVzE1DU++nKsN0WVe1CWcGZsX3i7
e5tFrQaDpIg/TI0pSdgzl9P4xBBaZbphrnFOEPCHjbpG8dyhEJurM7kNSuY3k08xbo7QZyy1ZTiWqzz7cQ7QQcFmJ4FdASZqGr3p
NeBp/yxmNGpIoYIe/DPeIbYLCmq1PNNX8f0m7RDPsabvhWcTc+9E6ufEnBuB2dTtKABnBAHGE3q0oAlgBtxB5WLY7uk0PlGps59f
pIS+woCOPVgvMhRe9yZEKTufVhEb2n2jP3aHnOMcB3uQshG0NzjR3PLZByLCUemaplSoD2jQVLaY913bOtjlgsK7VxlL+w3Ab/H/
K427Bgg7JIfDBjirngRnuCbZ7HnNxUIZwJiT3rlJk5akc3g06ONHIFJtop0M2zXUSKfmqvky1I+Buc8fqbFnnaH4UNs5JriJGJ4g
fv4edHvAW9kEAQwCRzJclNLXq26rGft/DnfMOR8oMm2DRb3Zs4nvojWpsc8kATbQ8TZ2Syu+yVFqRbP7Xki0Sxqle8uDn7LodbEz
z3J9xU6SFHtHe0QO3hWtNaMAQDAoID7kd1Gcr4Pwfhcs360ZrC45+MEYw6AkwxoeMGhixltwwfwPVQZx+erNTE0vKRkqwEgAdbQY
+v10/pT5o5RGhjEvelQ11X/69D0szqE/l0t/fdy4PhWtT6KcsRMgvXJyW9pCqV8E+nMCKoOQRkoTKvI5Zqn+Y4NAd4G3FXxG9RHY
Fd4rq8oS5xMrrrkt/UIxvsnKQxeGoaN30n21kae/hvEzw15bip/uc2U/bMzz/cqvHjolfkRmBxUsralePpU2YhjrcW6kMaJlQxow
UVQXXqRmRJIHpvnrxuuEBu7icSbMi2Cg08I/KZMkq1S+1Sen83putKkPk+RFCB/94NIYH6Q8spBzjaXT30evj2uTH46incUIwo5t
6rAqnVwc9cQfzUAq8y1ODoG9CQdZV6loBBWH3DRQUb6fvlRnVKvIEf1XlL/YTha2bvUYUrVszUheDD29q01VmWn9/LLbbc6HUIXR
fu9t6UqvWt04F2LhoXGyquD8wdrXHMnzZeOjc4bPfHYJhijzUpCdBk8Wka0eXwqMTFxROfad32LFslztlsK1x0hplZPZK/XVgD/i
5gHAkzIvuKlV4xo8aPN9R4xjVyDZQNvUzHtDzWwPst5UqoSqOBtjRGAWkn+E0+McJqs7Ivu4NDjexY/W0ejmn4lpXPtPppoj5JLN
TRloogWoZpb16pyxUBYZW1R+jd0idInsgEVjskb1dXnpvD2GBOT039BynurhWVLwyAzLJkbR11whpdnfXAqIC+hUPue0A51tbUmy
VNEBYqJG99dKpUpGeOuuSQWgxeYG7RRm6lA7fJ54NPZPrll0Q6FvrFEiRFVTDaQUeaY28dCRRMU+4r74EwZTtoVarDWZYauDUBnp
FGLyI730tb1DJD0GXYs5epCL+mmbenpRAkOzJhbXrt15P6g9PFHoZhc+wHJaCwHiYr6sTTRFp1eUSoeG5KfdBA6qUOE6+qMKghJv
Ce94ZJrjk3bX2trbFXW+VOdSgXtD6le1lr839KQxmcqkO5GvZO5AZq1RH7Zvr30fYTChdWUZfXgeA+g5SyJl5zdGup3PdB2yqfxU
qgD+ZrTs1/Lj9s+xW/uOV9MytRI8mvYTBEMxHmqPXaStQPpvhoKssg6yzWCSU19+bnQ5pFnkBlehPV2wscWA1NCpn3c2V9Je09zb
zWQM28WsjNC77yzUWCVFR2XasH3TT6a2Ogrj4KE70bRgfz0QOghyUYuNAfSouj26MWIIUimQ/W/ccf1tbW6ukhtKElGAetWTSg9G
DvD/jfw5e5XwnL1foiYIzMCrodm+dn6b2L38qYCt+3mcas1O6ypIuf0Qxyg9OGYJW3MKSj9/h1zPZxxLZ8ngah7buzHX6NO6in2R
sQMKuPiqrfEr6dForjShudFfrOT7c7m2zJ3331KS0oyWgy+px5z5ospfHEC7rMuJbNjLEx3Hhgaojrb1vk+FaOAmHH4i4ivugfnb
oua1NVoGJmDA8RontNEIEpLLSzrugz0CRyB69/q1urXz5XHWTmOlhqr8Y1+0VaRYjretod8Bn3LA9HoNqWZAR3Lo92vsIPRPaBer
LrkRIGWdM+9SXGU7QDJkL+YQm048Ee63gk6zztTzHJ4B/QfyxLAcbKmeMOMz0fv07/CYa+t0rg3BcWBzofDCZVAdFsYGgNrpKy4u
hRjac/t/qcB5o9VjWj8JbtAwE7QY80X/1qC7Lx9tFK8WvFz66702iQyNMbTjKqR5NN0Y5o5jXk7eRg4OeXtR8t5Uj6K+62lAhgzo
eNG9rSHR8K+w5GyNn4A3pszY0kFKnS+BORIoC8aaJjRcnKVhfO3YPuAZUZK7sxoadk12I5uL83y7FPMSeFvPxXh+2mJ+USaQ81PJ
ee54NaLQtSGPT4TKSn+xjm6bDo+fBSxzkZ6QmbOevY3sxp+NStcErtiInYL1qLzbdaLC7MQkPIX9gFWThzlehLB0dHsyFguqyCPx
i/kCztyAesUo9XIsqIZ491IS5xi8twn7PmH5UwEKs764zbQfcaONxOkJ8A5wNaVlPWFIN79K+3u352HkT3kaokb+VJftu14+V/ES
BPDNiPvd+w+NFRSm0N0xG3daEX2LXAGpCRAUM7TJrtA4uWh5czBaswp/xcJousr2FMrUkezpIV4qQBedaKcBq7SpziUcUXhDSQ2u
bcxhVjVE65OBeZtRgfOwgo9gvj76/l/sNGQv0AFkvsIeqPE5f01eJmOoVxUHebHB7WUz0yBif1y85CZ+2YWVcNQ0hrjs6KH/ohCj
5zUmCqTecjfhjFyzOtxp8PHM0fX9kyhiQSjbcZaGYDIqqB5Odxmukqk4Emj7FsJjEQNdynDdf94Yf4HzmbrehPDnO49ftdmWBLj4
DiwGYi2nAoXi0FSAk0oYp2EGEcesp6lmUDPdZp6OqChDbYbq4AbZQLR30Ayf0qYzxpQ099q3xW/QwhUZtHWB5F4g6ccHMvNWn3tg
mGPfcWTQzW61Fp/9LReoheTiSzeccjPfUfK4TozlNKxse43OawtroXHUONUYu7DOI/HgVSGhhLnDtJoUZmZBNMfZ4c48ot8RWq5Q
NV6pFMl8E3MjTKpg6NOsYSqQLb7e0K8wf97gWPfAYzNOGEJasKGU20XTlT9BS/10dRiYGPnfph3DtXz/QermIuIsLytE9ogzBoLd
XwAd7Rwq0yw5QRJ4+PR3acFgbYaMndZwqqJi+uY5/ZJ4sWljwoIl4tsavw7HDryM2Jfeu1Sml5/Sxm+9n97BWXM0R7MG0xRBNI2j
Lxw5+nRi7AoESuSaO6ljolbzKuwVu7ECvor6MUc7X/heeGA2vSqk3K179D3BaxN/EPiYaDlMILTRDJv6McEkeYs0ZR8SSgUqtkcV
76e0zq94QRx8t2aIr7TJ/oNvFqsEc3fsnVybPxeIbKShHxpawJYjsV9friS0Z7VzWo+uS01biWpyhuQ+ghVJQJIxTBYs1RLdaznD
WnlVDZO2FBF4Aoojfn14RzzfFUFT59IO1iDMguiJARx12x9r38gJL5V4UKi18+geol+AT2acCwmThx93WiY1jVk858mk3iv28LY3
rFQc0arEQ0mqMt6l8d9GQoUOhhmvY1LmJBr142DGUyFk7z4ZoblrApU+B64vpffkn6DVLzA4IaTnFX1ovg5+hXMYVnldoY/z0XQ4
7/Y3CezxxBxK/yvGvmdMneewAyE86RgJXO6EDhb3xX8LsLcGXnnU1KnBmnM2Un731ZimgaRp4NF+q5eE7scEKdGqyBYWGvtY879L
k+Hn68ufNFm09wOBYgjMrnDeaJUXGC47pq3dWh+EtZ3mULkuO6iY/ghg0nHdgvniot80Tk7YEjVR1fCyBMEWCFBGbU8qkl8wW32F
OBugza3ZUulSqlFO3q95snj4HvbBVQIk2S2kkvmWTFCRJJ4XDYyqDCQP6Cn4gb51fBXa8jspvdGmzVGvyQid6lSgTF13I58Ch18T
gowDJa2xDeS2YcVMNBAzZo6hiY9gayF0oTttPiH7ivkZ7RUMV9hRt8xTvAbfkzTwgSO9HOcTj7Q4nagwgjZgRR72sxv+Xvq6vqIm
k8Ozf1/9ixGr5aipbxEHGYlhhZW6zxVsWl7Jn7IPh4naf9BCrz0hWk/aY3wE5yT+pKRAz0nLb/5+KFeo/87Mx9naGGgq5sGTnArF
swYZohl8BmaAoXKFut55rsGAT8U/PLHKjpkQW/8Y/ro/eFHaAGR/P9W3pJ2s7F7pF35pf7yAcQyfOlxdAV+hhJnpcw68NzH7rf7B
s424oIWf98/y2SmMyYcc/061gG/dBTFaE/9N8+Apf1k48fxAbpMM33nRfffvavHDdTptJKHm4KihEaDv9tC9O1hYtz39QeD4GZog
Fw2Jmw0DhRgA7btN4NH01fKh0w+jhj+EuPeBUsj4xc+BwbbDDhrDoxsHhtkq0/5Rs1Tm9sAul2py4BFn4BtqC1Xe5hMutRq6f9tW
/1bMQZ1W0E7k5h01AY5bmF+tqWqdktH3tg9v4/Ul4pIFZxxAgBV+Lc1LuMt1tinK4XE6mqr6cxUMa5N6KZfGPeMGh+WHg2LfdiJH
FGJIfkCxzMOQ2o+TLBI4MW3UxRlIPX0zQufL/GkQy9vBclMbcNtE6Ma3w6fQ9t+HyXCiYMK5IaWx8b8U5oinCl//o80ODVOVCqOY
OjQYw1CWCPe/C/Grd6UnGRs5fDz9uKlYobXH+6djaECqRVOEEkF9VeKu0N2Fcg2nmG6obsLomDMMZo3vH6177cA5qQx9wi750EZN
gj3FxGlzO3lSH93qoaniIA95iPrrgswGK0aNEjlC7u1GqF11x5hc6e7GRSCHBgBymj//hAjSdIICBdVe7VVUjj6jcBfXXXEnSc7/
X2rN8BqOeBEY+wx1vnBSSsZvhCi4GXRrONio7HAkMdHqdy79m+Bv8P7h4SeE+7KMvUG0sd4qzWTvp4riStP/5dpbRAyvKzAzWp/M
Jm7h5yAxPIJ5xsvT+89otFTKtlJN+L8NMNUj83XkJ10dQ4A1fCM1b//WJAJTRo4Z73d0raM1TsYKxgzLACEkMd2kvgGdCwfikqFh
OiP9JJX/o8CMefs6ZUR8JXHOrQKmkTmzoYcCL+8IwXH41tIK4PDV8X8zChTD3mfXKK7NNSsOT5NCSEFXEkAjebYyc+hVd9BCAqHK
emBRcY92/Y3I9oQFxLPCVu9EyD0xV9pcM81TntS4vbRg5ZZ1sycacYi397yefsRBvVrP+V6DlchLXGV4qrP7fw70+9pKZgBELyT9
IVNHFwpW18JMA8kupbNe+Aa9ODDbmOTcUle3VWRXjBaaS8D9299n5lXCfBz68H0XhAog2WNZzGg1HPOGp38uNeJvR0/Lo4X+/w8L
4Clo/PKGzOHM/ZHdYI9r0K66FRfu4fwcSaq6QKfUNvFxIDcRC4sfwIdbst1c0QvphvCOmB2Z7mdGMs5gE+ERybhI1MDwKv2mvqH6
PU3+2h510bgfy2aGiBhbwTeqab2BowNcC5xaog0rIlcmW7TWRYaLWK/ov4VWUeKb138Xz7XYkopX2mFJ4pCfpwh2Yd+gQtUGVhT5
xX28YGk6vTJsoKj4jM5LFcEB+s21dv6RWTdGcuf/eCL2759m8pZHU0c3HDxh4MElkPwluJ0seCscgX1qK0pBT0fzVeH/hRxPQpWb
Pj8IWE7nekrpcR4MdHHcm+Hk+6p3WshpRZpWvvnX0le0Z7s8Pm6mfj1oLjcoRuum9IThiTVRAiO9bhgAIT3129u0r/DJiJkfCVrv
aixLNrvE+dv+Jv7DTHAe9AkYdBWep+MNkl1BnQjYuaT55nMH9S7GO5rh3dneGbg+EcWqCSJfVhi+ZziXuSYeTz32kNdR4fv0wYwR
C1s7Wue0dYKtzaUQelxaffy/JhMI96oKmFCeEDW5ucKJOu1CeMEIbeHYlZ6NJkNm6ZqJcYWIYvxbQYvMQzoqQM4xGMUKktjJIRe4
IXCVTUwN7iaJrZEAqj5JNJcQsLJSHZ+qHVfmw8QT5nE8tXNJfgp2R79G2sUH/E20kPeUim4p8EPW8ARRxl463Kgjl0lXYiLS0452
syrYtKvn/wmpE+P1BGlFlq2IN17coVzb1oDEjOZ8ATBL9/VxVlb/IwnHM1prpHlHjQhWb7EsAalHkEQxrPqeLQtrCQ+lGdBnMB53
xI5G8HxVmqGEYAulNSl0HEDIBqP8kmzB56Copz0HijnAhR+VyE6+0bNB8f0E4cngU1iEmnGjPxIEnzcjdOxTTY4OPd4v7+n5pm9I
B2zDvavM6yCE01frqhCU380u8yJ9Hqr9OG7DJ/y4Dj+Nev4SQLVzy+OPbXH/rT+jIW8/KsnmTh31hu3CLgwC3+BghRINUYAEyD0U
hkhqjggLQCh72EiEGE+9s39AVl3L1UBeaE9OWoXwZTWYDQ/ykkWEboHO+bEMmhJy4clXbXCp6posBmDKTduUuvjsd3H8zj+tvyEP
+slulJUeDzcmgv8ulkDMcNP0x8Tp3+/o+4onVKDhpgcrsy19kM3etuEkP734Spy+pDqF3xRZBjgIbIawu1V6cGuEV/w3vDnNkzso
mNsNDUrOZs6x1oCEaCmBnk/pOLsUXm6ZljZ3iotPk1jixaM3wxKs7ee2XHsOkZAByIivk37Uw1zSdqRhagCA2I0CdfIlq7TNOn+t
VmwUU78W0bj3ihg69M+5J6kzKLSSoZizhyNEV+cHtg41nlFH/DOibjOkYOZ7tao2LloPqwsScKCZsnsnmARKlS8/OXgF6fW/LHeb
jg6x69+0ZQb3Xe4QW3SF4F6woyka4yJBB947qW0f+wKuTaPS+rRGlFPIrcVvOi7PbWnPAvbslT7tgFGE4LZhtouZqIDtxdVbQQcF
Qn+qboBwBJ76ylNv/7AVx4/jGVHZLbE5FdbPw3/ebM/xzCSzjOFBDoRvdZ9R+1VC52WyDKHAtd0gUQ+ERlqRJ6CD8GySJ2RICry8
+qQzgtYY8yr17/kgskXRpfSA6u0c4zGl8D7RHZc7z0RAlZ0lH6M/BZQz9qCQ7KQCAUl5uCtwa6H1IB17DrS0ftDM8Iynt+LZPUzI
fKf9W6nsINB9DXA8lu6MOQcPcDdPwte+ABGsOatzGNLM8UnNx0Yhzy1Tn9zLFbbevuQridQ9fWwmCbJjhWhxBASEC7MXMptArUXG
hze/GU0tapZ2grsCJHpxD9h8/SFQo78VdTXWEmROZMl/wwPqb/wrdfohUOqD6LFC6on2GoHm4P404bkp0AtK2Tx8wtPOezQx0He8
SOb5e2vm1P/YAPemBRZPfvWOmkWgvqahIh90nG+S23QHvkp2eWxeH/ox98PRyisVas3KcjdQNGNBZt1djDRBRRLSqR34QiLqU8QA
Bn+A3oeoGGmK6s0iUe17QP0lS93eD8iEvEk4OmQwraxPW7y05sR9SZpXsVEILNpI/Oau6WHY2ZupSnM9+CKInXVZEbA8lrCZT2Kw
8r6zmyV2SpHE1anwajAIASGtKxAC52uS56fX4vHl09sqLJbMQieWijLHVcB1HkI2B01qqZVs6S8vi6KEfjpoxlJW0E0rMNkZfbux
4p4rYDhTY6VaCg79UXSHthMPNRNuy530EHQ48PDDduxFVcg2q6tZjo+wVgebCZSeQN0wOGtRyyslSHi6wmPVgpzE1gG1Z3c59Log
dO1YjKhbBVeA/8BaFJ7X90KkjmRuzgbvNiBGL/5od2U/+DuOjTearW+h/wtAv8LOmX0wPiP6hZwJdB51IktgIo0rVHuYzub5PEhE
s2BvEllPnLn8DUq3f+2F/sv1rd0UNz57zHjkUgkwwmbapDalf3cFTz+IUFFyUbovKReBKe/Q8+VJ47gvSaJsNDAo4i4bhHfFQpTG
OUOMek3UaTG0Kibs95JgniFNtTeaHeLoe48yTVV9wklGAt+CrGDfRQnj1L3jVaUGpKsyRmqiaQc5kiPmZHYfZ1Q+DGBr7ZWowR9J
KH47XhIvK+WV3PDRDdiNEWbuwj+LHL1c08du4sn7wVckrGp9yfLRgYz88PXMi+JQe7Hbf3W62vKPCmTXQbPCQDC1J+p3FqcJHxTj
FV2Uv7c8Gn0YA4EqbIHT+pNzy6L0zVQJqxJb+flAJDk2aEceLz7BDhltyw1MgqqBZRAo232NfFL+vlq6POf5TWdPpP6hoR9BKgLs
TrbXhEjVWzAsK3/3wC0zszOKMAQVG6062dw7t4FtRDJIZcZF4DOny+UrVYnWJheJgS+e8Ro1qnre/Ge+w3mfCC4UBq5R4f2DAHnn
uL9W4Vux5uRlDYQYujlBx+FttfeBi4g04PLko0iUpAcx1ee9iQ/yau7bl8MQRYRHtg0phjHjCrKJoR0LOjLAVcRiBeLWJIKZvDcc
LEqJcq2YcO6mPPGmWrL0GbOG1Bu+XYOTO2E6mRWGA3CAon+4rxpzb9s2wggmdP7qX5NQX7+NUrkya23op4bpC/LttljZ+CKuMWrY
2pl7wixPxY8ef0gGUoT9+/KSLYezRsskk1pbTYHk4WMC3cSnLcEv/qjKz1aQa0JmGICfCoialg55Jk9tK122G7hCI0PgAUu+1dO1
7GnntvPuvKdVBQzFBWOP0VApNUXzAYrLCvfJNhgS2m9M3CJ6hJuubNcCHVY6JWL04A9V13zNYN2uvozRWGLG3zMfgcjrVSF3emac
Oh2JTenrdmjCM4stJgLF+jMg0fFGVmXHXxfgubf4dj+LMTnTac9cl4kN7CD18k8XE0zg/1S4LqlEppEJ/yFC7c6WsWAH0/xA5VUM
okhAxobUlTyzP6EpF882Hlukw+m7gIvNkkQN76xLKdOv4fe6W7Xth/2JqsEQ3/LMS2Dcfizywynu1zzssYRrfDKO63CXrSWxJGrN
liTPdm+YQB7Nek2QF8HErqfXF5/K7LEOqmjykUu1odQ/j0jEkv3UK0/nZyAFh+vrpPJf+6i8wilJMxUjh1KWoDTVJwefd1HEXABT
iZ9Pgd2/WNwZfFtHzQtF1VrgylB3JiNNBjghZRzkRvu3T6e5K8ZHp1gd7SRQfD2h9Zy8TF8JqIC0j3Ne+LBLZc8fDAjvxNMUkNAR
z7t2iPEO9Ye3dLSt/99dq+1yTLySqT203OwsOPLrffsLD3V5PEyKB2JM/gEyv0F7/HjgDUXWfgrNQexOIjlYflrQVgXVVTHMAEDD
rFiLQG1mIHJ8VSIj/FajTpO6VoaUl9QchYNpU9x1cQOOU6t5aGfins/8uClxXQhUIlBLaKtSR9owY3dGN7ZF7udHDrr+4Rpxh9wh
0rTx88tBULtWMVhN4SvjCljZxRChMvmL6dbT0BIHKee939kT2uBr4PaecrZkU10MPDHjm1qTxtQHOfL+C26StIGA2lBCZ0/dLv9g
2YewyhLPD3kjtuNpPH+AA6mwNt+X/hmenNXcf3L+R9OB1W6cdegzuLI8hCxPG+jJCj0XbE6t/nNCagP1pKuFeaShrrVLg0lRHu0P
l14QL4inlAvFCGmT+D8oiaRA/UuW+aCPdEIoWIBaH2lo5/KT7tCQUmKB5gLH9/iHplzzs7iB0WtnFAfYjQy8yzPOIZy08b75pNJ7
sAoxlWEvFkBCFY9VHt1KKbnMnwQOyJ/cQziZ48nVpvn10HAo17VVq3btMnPOjEqd80i41tqfXP3NpR1v0cFSNmuEF7DoR1dch+NQ
g4GUsoIi19s9oR7YgboYIoQoNJr3KHRdY2cWNDtKls3ork6Mg8L4QBAzOW674yogwMb4D57zDsUZTTYGn6SLXb7rElDlaMNWDGda
WchHMFxNic8iwH5cVaKnBT0xPjeUVxWSLYT7O9qXs5nCQCRroGIkcKn9wDo743W0X4Ig8DTbUUIEECxM+ah1/hbSAOJtXDutP1wn
CYDlKddZTFAI/OJSG1uKm3p5aNLvqNKeqHGf9XJ10nADj6gM6MK87qjKmdvTp3G2rk8aIEIycPybGJMqqrKzjDfQh5jEbMsQXh6l
GClSHFz+ertqOr9fQJrg39LgYoFEqEZXRkyNlwxf2h1M4hX/kyITrtExd3z7Ybj6K+Z/0nUDJeuvJzDDG2Qbni8HjpQ6nnFNOg1e
I3nSquW/De8yi5aUI9YfRQZFAUczQ2VtvinidfEhhMbv4jBnk2lXkA6oup5OJDWpUH4iEd4NYApRB6Xx3j0Hs1HWP/NvnPKbWTY5
XX+N2nTuMfE6fdhmBebdsl5OKx+cUx9Ppkxg7mSUEGMNKUvyY9nd9uIIhFLFI37mi8hfRX/2RKrIV8zavU1VqTbGTSVdTpGihik4
/W+TfeVmoYSXdvigGT12LjSWCjXTEFNlpNz/lKD/CNOIHZ4Fw4A4qEF9mFgXKp6D6vzfA+/fIWB/FbWw2OP3A40kptTNZ4mMS//K
NEkoWzDBTebptw9UTegclt1yF3zR4bgQ5aKI1uOuK93BRCCfj3l2tICxTkkRK7Qr261DPlZ1voanujR0U+qjW0qU3OD1EoTN8/0m
H6qH+J2M3dRXEz6HS1IMyr+nLRSqg6XhS8h0pqocsSZmu9s+Vo8KWgANw3jVhZ2VVD0QgVVcOE6ajCn8ewzqXgmtDGfNSRar6uFj
6jTP2WryFnTRqmYM6Cl9WmXfMbIZ5kc0IsVfyIqNgE9UWpnYroPUzLi5B4bUq1jIKPX8+PYneuLXlEkbo0bnT/OuzS95eqHEKpsT
8uXVjgaLEgSEHsWeXgBh9D115ssMxw6yFdLxN4nxMrKWS8+yJbVqyr9AOYune/NXoAJoqZzEehQbuF+UP0BQ4iA1++s3n+qzbbjM
ZGrpLHRNIVZKyqqeukzTOfqCsZYRafg/KcipZsxWn15zpv9qaAtq9k++uPAULQQPmv9ivej0TPo5PDbAYjLh66M98KWtbR3ZjkHj
P3nwAgnQmUCBuD8PKtv4I486TRDmCeT3VhxX+in7Zd4Rd0NF2fBMO31uku74tFqT4YMjRyCnuI2UYqOFv0da2OfgwHgdVfTwLZDx
C4F72rSP/349+b1LYGIRGufqPF+H5MXnGVD+Yn/GsXDjNbA7PXOSl2kI5eHg6HM3ehTQZX79e6SiRjME4+OKorX2rKILfvtWj8oX
EkJ+tvWMbaDSSF565PMx1nDnMUHCD7fJO/mwoar7cLjtU+U5z2PgWTTQ4f2c+mNIU+dlwCvghJMWkDuoiLR+OFQcVbzn1MqSt+ck
eTEncDkziF85AigCT+q7GFHlKa/MCt0kdH9E6SmQl8NzY+CkIFp+C27TWHe4zm6/FPSUdN035HzkahgRBRahskp9sp2bY8VX9kYl
7iLCEtgGoK2GUhHQaCPgVpj+vXL7lYwGwCz+e4xGl+wh0m1256YJweR7Qw+bUNxhjLcDtCHBzwReVDVv9Jzmh9iY/N1Vl8BoHyGD
O55yZCPn0ZNAtIaFgsy6RXbt4yfp4MYhOfUdY+dUZaNRSQMbDecdfxLgq0GfjSHW187IGriYAmEPiWNi9onxIV77guELo9NgkpqE
o5DG0sUzim2hpbk59iu91Xmu2z385pBlBPtZQQ0yXGlHVw6L7z1CY+UHRGCSEXTdf4KBtP/oHdNkn4R0WI0PEVm/niHjvCN4479K
ZuryzQwOJVQx1vZEeJM98+Ti4oD77wo9erQsKexJAE5OTUSrWdkBXjkGY1VscyxKNqksNEwVSjkiaII7CmwjJuBMXw9q6R5m/6TL
FqFu6/+9wApWvsnD8zQ0Weo06B9VMTPsS4/W5dZYgmLZbUNL2zgMf6L01UeoTnt2Q/LDJRmQeGi745qwM1aboB8vHcePIPcG7pld
RnsTRpeb15ltdGPi8JeCRhLfZQPAEPrTarP6MG3J+mLpbR6GaLPpi1+D+vrERbW4WuFKeirrgCbnfWqHFbKlKqSPoBweHqWmFaKC
PzYXG1SQCB4QpdadjG6Bx20FPiIO5EuEQwKvrwU/vSrQ8DC9dQwfL/QX92vj4ldrvxQ5hbqPQKVn6xPYFfX6IFVu13+FYVxBBrSU
alBQMfxbEymFBCnczKIXNm3yrcf3NXoVwUwND3lWVZgODpluT6Hv+QHPmhqtpPHNIdV1uuuzD2/t4eSNku1dI6XBP109MBpFtDQp
q08ttm9p9lc1MEKuRQBkAd12jzPAKT1zWJp/hlkXl0hAfHVCs645V4DhoD3EIqi++SSEcXj95+P4UiBq5BUMhsqjpTM5c+m9+7vK
bC8E5GO2hspwDZEeLL53CGDxBXzya+Qm5J4TX4/Gxz41nkqpsiE0AghDitpQ4iwmxbxF88lTGVrCUXT1YFw6FKJKHE9prboAjZ6G
8bbAiBnSNWmnm6c0oa92e9/x1ITiBQtR4qu9qZTm8MIAHtsQWwfUPWCzrC8FLYWWNKFaRU7Phk+KgKQcJ1ARu6eWVKQ8jPSANoKB
89g1OYsGjxW6bVZwD09dKoldknDdfj+8iO3Q4RvxmsyBrgdXslw1/Z8rfAjEAkmpgqoedw3i2msTaJJGOH33ChfZHzenW4+i0thc
9ACeLwtxvmjGs8zjLeJXCmAQ9EfPZgqnKYzPcs3qbIbJBjM7t1CCZ4GPOXnbhP05nEVcQR8b6Nj4SNa29fLT3N14zGkKLHFvCbgH
Spndpio5D5oAsVwurxA7QQJ9H7CGPtbqrNUfA+x5e4zyUz5lavJfQLHG81DzhrCCBWF1aV5jc43qG4Uasla6R0kWnZxOEdgKr+/K
e3yDxPBxboqPbRPqmpUJ4kZtJDibbVsRpph5yV4b9Z8lG8WACxQDnHblb7ANJfy8YZeSGLNhLQhRo/ks9cOcCvbX10uZbVTJat2L
wFCglkv2q7IcKdluRSVfHLGxAg+vsdex/l/1eK/QOpONcQPu4zqJzmZNy57k1EULw1vHQ8h3/3utSUrvWXtU9qvlHtH/sO888QjD
+ycGSlLmvG90VbiwEFTDAcgxlPp3ngsiY6OipPYHubHUCbqS9dYETguE8VtJ6R+dTYJrO+c1lP3LW3rUDHR3h3ZgINpkN3fORwYx
Cq0jQYSNy+KJBUrJWQaB3XTN4S30QGYi6QnL0hJtprJQxuLwPARz6u+DSCqqzktim8WAHKyqUel3wt7aNY+/qGFRfDm277xqjHsj
FWkKr0dQ5vzz3AOTawoduB0mBE4dW4QZ/4A30eFR1fLpROca1uVKOgGNEiau5wtH6h9/NWdhuEF4L0OTRoZbLiFmnLPVM+LlFfOt
SCeQloOeUTKZxs3hMxVqHKI6MI34kiLkf3q3CF3eQ/LsVueQjqAuKEPlZwtI6FYXRkKOJmiw3YS4q9uatPwFRUXo3oL3sXJzafYW
6KBNLhSQfM6MCpBC2OhAR6wy3LUEJtZv6jYtRBjKK6OHOUW/lnqupdHgcEgv0DPVLvOrhPb/BTyjjJdlbrDNOyIio/bo2LZbeTbt
h4l/stLkREqHiz/B63hkEWZ7ocJ8wiyvpfC34Vxk2KUEY5mzn74QI1kD3TTEvyTNJfMYXZ10BoV0RwUSaIi2NxRAk2v2xaamfICH
lwxVi1PXsjy7k5qlFIl01PAp8tY4pCWw+tgX7L59+8bw7IYOlG3jfNxvzTTEsN1xUvBaSYDTaFp7V3uOW6Xx474IlFcv2OlW+kj0
aoBHwxTyOzewvntaNgnvvBUqltsTkMBR409p2uSm2Jg3ziRfeD+rA25nZu4ESZGxUYKhT6mlsaJ/hbQgB8cXusOsjxfkO0fkwQS5
cPALTMsuFLs5qAH5/FYbeN3ghB2bb9kfCKxoX/xRmXC5AInvwc7jbAHrSa/dfR7xzwGGZbkXYx7My4wW955X3K5/0TI5VxmhoNYE
duerLJyUEeRzMPZ9cZYnh04t0RiGBSs7eeZ+CDoun6DACHgyrBh/+phuYLDpY6k2oOYDg8NA6mxEq8ZgKc2m17i7HOj+yr7/Zex4
X5EcWSIZ3LvwfEaoXtLjRJ+405EsjfGhpZVDoB9PKKqtSirw7tKuey46ChAHu5P+qTdNLB8YtE9s9I6cIg5igzdec6bxg54ob1Tr
seM/gNUzIjAUgwZSF+zloa2RcLXzO8zIWbYCJs4otuhmgPmNICDa15KJEK24TYf5oQ4lEjXmJQSwQ5oL61f3719vQGON49aAruaJ
1X9QYFqDf8fyDjR4wzaIYYsMiTucYXuEdKRsgXR6txRM5oCXqKbjBHUs02385scHKry8is2dTsLBxX8FNb8YO5H2HHC1BnlDSc2B
uM2W9Jmpd4+1y0yZ0hXC4rB9H8PtgUyf23tPTc1sw6JtLOx/3+so640fqRMOu7zZaxHZmcauROkLhseGq4o9OhdVCjOHrbHTCcSk
r7cRi7u34domwi21fz5VlC3Dctmw7O5TL/6MZgsj51ly76cxk2ribNTFcxwX5+Mo1waxgisX29vGCXWECPAVMWD6mE8JvBn7Q6k/
uj7J9nifITF/2yvllSKh1TSP+CAqRkHQm4bv1GEbY9DulVvV4G8egZRzYMXjLkXqVrBxsrwKmkoDfjlkSFBWRqdPEqpcjwoIJGnT
7cm3QKufl1SbGTriR8hfbAXT53M9eppQbMyIgRGkKXSP8y3ey04Zf+EYZ052Dph/vpQE3tlCPF8jNsS5OANVETg+pLTxd8WwqOxh
izFG3v8rSc6zSD8p3NStLXFZFp7HvJW0hoEMD+gjHwaIRVP0v7lk7oP2AVaIgBUYh2Y5+81zWxvah6YsmzTBeZ1fkPIAh3/E/q+B
LLZBlNO9QS7xHRc0FPSPNXqJ2E0KDmzklohv5blvWMrJ8YBrEuHXhJ4qLgk9RHEO0QiVYShDpcgoLyK/dWwaKc8P07FkC30qbgpa
34qmIpEpkB45cvu78GVLffnxi7c91k1/dxaDRiVX6FrZIfUkxgrmLFxk3HzHcF79u6UPQIN3c/Lu1qoy1vrWFtCA+RJ8+VGm1O44
5K5+gIVO8UCc9H+GTXTmkKLuZk8EOOgQb90aHQq4jI/QF6VM0bkLkVdMgr1qbvTnG5Z4kV3RkHbFQgXxTAsrs1XJLalx40QlAEQ4
Tmw1BA6F8//5U9sskiaksdoHwpaDYiIb+SpIM7P4D6S8UvMp3Nt0ykIk66QSosa4O4HQec4uCf7tKe6T+19tZhnVvpiGtA7s66kk
kHsU98mrzAGv6ICndQ488f+yVzFQOuzRQeHuakM2W5b69+85F7kUrUi7izgnhAYkH4ceMHgUCaBMKzlilIzPr7BPiDo/tpsO+PZh
ssNCq+A6Q/9+xUbz9/ysa/INZEH2wo2RWQTVw2M8zSP2rpEulCaFLgm0Z3dz5I12vjbfni72rSbD3fYG0SIzb/K9fvTQMtgA3cIU
pjvRFrwD15wlI40VaT2t5jIlVtc0gGKlOtIFhNZusu/WN3tTnl/n0gciQdC2EyDeeGZy0nOxDMR108uhjosFhRrRYVFZ/fKg6CdY
YTcn3gkpsWHHeI/17Qu3gFgXelYC95khCgP3QtqS/5OudKhokZJNAn6UlbPvWkRWEG21tBaBbQNUSV0L3jsNAdp3Vcz4YFizrUj0
uIi6Itl7RJPw2JV7Fd4zusYfL+jU/G42d3hNNmkfJKUvOCQ7WfSkWJ9O3rCZe3dLF31A/ri1UUspo4KccIc7UcHVxZ5sJGKaouxy
EsT/M/yEtTZpLGDPYOwbz4B+KeN+TioAJGfEaToUQA7qkD0PZpQYAj5H5pdOp3nVNCMbZi0ggw1aqOT/m+gxj6nh+2BeW/8kmfx0
qcGyjZ+Gmw/OSBFa5gKK+mTMgtpcXLqGFCN2tctDOQzSAcM9nMENqup3jvLPdsW4sxS/ASlHeW8DJMZZ67EuE8xzYA3nFxRYT6kZ
FB4CA38APhMjH5Qfd8yD853f/RHJ7JdmXAcHuubBcd22W2Puw0mnbHKknRKh7qOMLt7qwUhnUDrEQx7pYFbprOQD4FkNV/2ZsXXc
hz9wbpNuDjTkyRl5Hcngdj1Ac/Onw7slN1EgFLcomqfHXJ3ak5rXlup2Mz1dE1ajBxJu5EKr6m0qX9Bc76R6V0nMZmcfWU81ZnuL
D/i6QjHg73oxlBXzGruCDVjl3hmPtTKdin41Y/hmwxbwvkODKCFMKSb3Wx95T0tovYVtMqnvspt8X8TcHd1l/Vbm7uclqDUjIker
9kGPBZd2xq9NhxYISWXFscL0g5wdStjGWsMTj8FQa9Cx1fWBU3DRZB0muYGajfGhMMHS5jxcxik4w8EdE2UQMJAZHC3M34DHd7qq
V8oQwX64wRcxZTZ19ZB/1e/UOwzUi9XPkngDXswqW26iAKrisXk/eZBROj+mgoG1avKtJqIlh4+Sm5g/MiPPI+fuWY8YmzfKclFn
qL9lO8eZDNQpmpvnQ7NixyGTuV3QS7clX/4sW/v/WdN0AjLoTtCWF1eH1K8Tq7aqXeuR0Us+OG0lnrclUeMbz2zTPnd7s5NgEcPO
uUDrX6RsYdOGDogGjqxbdGTy9/WI7BlUCrQ2eCXu/ZGe2/ONYI9kk75sueW2/1iPJtiYMJG5GCV9IeoClIiWJK5xslri6CBeUyG7
n1IJe3TPH/i/EKOnlTJ43wSoUBjEw+6miMeW92bbctye4Fup510c5pBO23cfhXs4laWErcwNDMM04cl1j98tUkp3X8Tadb723WZz
3FrcJFEaew2QI+o6btu/2J5Ywn6EKXnCLTZ+Cpan+qqkM9SSJeBOC9kuW0Hg/OYGBVmH3X/ACuqZYVxd+xHfVL5noaOVkuk5KrxQ
akQbTEGsf2FCP5UdiILyFyMd1MBCSJZd2WMvfCUS1ozF+5cItW2pFdCFa9qLoAI93jKZ//WWcO0CM+Jc35hnpFunQZU9L1NQJajn
s99aKH1cQ0nwudgRPM5c/JV6cZyfc2rD8j9n8290szOI4n5ifSoFps+YL/m84WFHd5Lo7yRwm8dOB5ByosJwSABHQAug5xn6l3TX
sX6QquHlKZLqbtXJmEwTdx0hezwdYP4tRnEb02yM9U5Q/uGntFISuTshv7BkdmTJf6Y0tW2tYfTo4RpJNniu6iHlaLyaNTHr4Wnh
UYX/jDLqgeFOk1+l+d+OGieQivSsIcgFAJCvyQJ2YPdl3EFPrilQnWvrxQN89ywLaa6NzyBk+nHEAr3ZBDJya0jNvoQTXSn5Jnxu
ef+lei3HgRVrALf6hvut9jb0YmqoK9PLfr4LopW1rOnKf5+Ob+6xZ2cpTXKTtyUPurw8qMxVFTTrilOvw1wjTFslTX7/OIUc1jGW
NDLp35rfFIPe5k95eHrbesZtPGXnorjffbKDe2/6/4cbGMddkxJAYJnE0C+iVjfJwpBAcUH0nBc+GLbMBRF4gQsJDB78/E7OmLBS
Vyn+Ud/4DOAP+C7jEFpjfuRd5mCTAlJStLfObtPHFHAxGaXn9LzLVwEDHnVsFK8djOvQHl4CNTEg5mjyxGgAJ9CsndV3fmu9Lset
354ennisVNTGoeyokmWfaFc3cxWmz1vqonb+JkQM1hSMKCRP7agsvuP/9QbaPTuy3tt2bjixsTZxsFn4iFyCZ+7k+s6jQTaivae/
t4rgtNbMQC0ajorUC1T27bmPIeJD7QSikwnDUIrJpOwU52N9RhGVnJTdT1TFTpYPN0NELgwQbbSsWNXsbOGv3j3NLMkaRU1+q73Q
B/uYZNgriemwt/bayeDuHu6nIFkY6KTwjB/2ALgnps4X0JY9vuXJTYJaGu7Uv8xZhr16dzZhKcxdT2GW1kBCAcU7pFW5aUqUzOUt
XJcYnVqfCQ2QUv/DjvSnGMH/XxDjbrLh6zkmCvgT66fmCYMRVapwMJaEyUrwKHZq+JZezqkhgLgJiq1qsa1RJLLKEL37rxZHsBFm
pESuI2W2PRMJNhpx+D4MSJ91MC+AzMOPcPGo3Cvy1jliEO7txVIvqplBg6c29GIvrTdjYG58vsQI6PpTGAvzGyRpaYhlBQ66X1yF
D8IbnRoyjkshaHE3/4YsWfRyYtMpnnUhBB+Ku7xv5ogVNlH4LEhMGPlI0WPZgv9tNt23SnGJOFGEYBiDsoL61ez8Knc4O1GIz+G8
DgLaSfhs/YAqkbbEMTzLbGApiSZ1s81R5RKTxOSNcxUhcPMeutFugV5aZN8wY91FZ+/RDys6TsmamI5D+zt93vBkPef34FdnfeMQ
DxDUuEKoDFjJxs2a8GyRWEU8SGfrFeSE0H+/yzNRFZAe/ai1d8KB5JfO8CwVM4ICwXoZA8Upuo/0fuTnTuopM3vPvBosppK2a/hA
YQf/Ire0Ebzn+kamwAKuIREAzQ5kUrjqbtMzq4sOp0BFNh17+Mz92FD4VRTalktChZIyGc/k0oMpzWFXGSk6dR4p+Vu11BYcKShW
1LAx2Fglwi2nFvCaYuOxJntSXa5TXyXS9deQlakgkACmu0Yjo17F6tkXc1mSzYVLuktZ5w+mpV2lIPlvskGOWEtXRLPtPtL5VEBj
9cuib2oa+CAEDNkDBWv0sHSgAY7mia4olmqaA9i7RoM3mbwRWuLlFm6G5RWFhprXYMTNditD3oGCpxbNmHkI7T2HW+wEtLc2u77M
KUVqxVr+6YAmriNxfvBbQtWuESDL+oc92a/b317CXGnl9zNUkB+5mfGjVItx0x6eaVErUwSrmTit2sFPfFKN3pUJYYUT8ra65X/X
M7AEqpr+OdZLW8AJ/3Plq2P5wdP0SNO1pbXHRHWUAURJ6pdQV0BopMBfqR6JpCVztsjWwFPvTcdYt+t6coD7HzuWMF9Jjx0Cj5dH
uQ6w9EKWm1A1V8v9Q3AQ6ciM9Htc+s4aLq44XZkfLszcLnBMFECONlitdmSkoXVixGLw3O+Z+NLVNMXqy1ser/KjmwtHGYMrc4G2
KjpaIqsnlf2bP0UaZ3bHbao1LPdnQcKEd4A0ScKFheB+4vjqbthEhqMu0Fd89380y8Gjwn+z/H60mMMhzSl0Y/XrSyKbwCSdeAZR
gg+ywsBigSeea578X7L946iwPMLGJ/nTkmXCFlXg2O4vsZmvGy/aVqBVqQ2tlkFJHxaWJUMFWY3DfE0dJThKE1/1ygw7xdXnggmR
DJhLGTUbG/N0kD/ABvG3C7tNB1HMx8as1kXRVcBZsmqVuc+6E7iRC4OnSZSimr/bJvXwg0IrJX9FfRd0+I2w0PXveiQPYNUu6Z9d
TBex5fjBMr5R+UvXMKO4mO8vJqYK5Y/QVTQR2ezwmh9sDTpDIm2Fj/EYITfG5F7e7KjdA8MYLof8ArMwI0HbJ3F11/vU/pggQRx5
1lq8MGNGgh4UDUgDzNdIO7y9A4hzgkEahmU5Xa3tebzNuN41tuGKV+nKxyJru3xzgkdLVrxTVz234oiO8Lps6wgHDnPAlxay7jFs
SAiVIWwi8eh9o7YSFGnyh2wSTQ+P+6s7fqWaGqPQtqPCsa94RMdkTbplynxGYTftpGMlJ77BRG1jV6OLXy63M7jUCkBhCXlBQB7Q
3KQtBx6fxHZZbv4dv3ETmI9LKGs3kEKjDSdEy6YC2VwbgODwpoQg/4WdwMOkU8a6gI3a41oGcSWyfT/PclonK5zdEXtUkDkvhPxG
tFHpYJQIHRhg68EjHjspT5FGgTfKBTC1mhxZEHMNl04R+mAilAaw47u8EdF7PF2GNJ/w3ROI1YHinQpGh2keEyKux3kU+x95Mvyx
mfxpHUGjBZCyGnXx2kvU6uQDed8iuWjasn0jMnp1CkcePRa1IKOssznf7DCIio6kBc3PtEpu4f2dHjUAyrTzmH1qPvThK3zeDigt
UWUySC2MGtfphUudnWQaxYrT7GoPCKAnoXi7WR3EHDRXoDBt0H9QO+JkuNqfS90kBK23uYV4PyF23KwLP987PfgI27/4bfFrCYyl
fXxbuqQW7lr2WH09tWBwIRju5g3hQfoFbP/TfoUWgzDABKu4utxnMkWGyYJmN7hhjp284YblFG8GCBWZ3Cgvtxx9bw8cILCClbW5
YxnhN1x56yrnvqYOz+QUW5TMDHdZ11AJN5hnUmmnR7w6ivBOxT/Wbu76eO/a+Rj6su3zD5GHi4KxmLis8YcRFmmeMxKweArjkDyY
VFeSexv2VM3DyqYUtGwB2VekqaEQN3JHMRrDQwb1Lx1BjGHJasOhbd8hrMGNDgZAmRbiFAfKe2IdznYEsgq1pyuKynfPMgpVJJsa
YEaZBa9n+MDWECJ9RqoH9oUo4fpAChIdUgJEu0tZbOqj42G6dx3BMI6dukem/tFlCdbubdjaQMUAnxB+BrfMo3RX0KL575pHsdYV
mWbW7ms6qsg53tnaUa9x6D6E8KqApytN6M82S4NL1qRpXgsHCTpm36y28D4dXeWqjtL0d9uWO76u/No6qMLcC9GSqIBUNBtqJmBz
ZPrQ7kRpyWKe0EoEv5Yff8rubuXe2sRCYuWCmLJMI0YGbqJS/inu0TskTLmgNPSdxrzdsksOD/a6C3/nkEc4u8Y/GGhf7HB4rdcx
d1nMZ7TPI/0t9T+tnaJOERd6DO2SgqwcfqaFcTXVgnN93YNlW9lCN79iNxFmyl2X7pvVvK5WXYXeOlJiZDZf4D7aovoqMZ2ZpPSu
HH9RMrIZpW9CzQrpAp4Qs9Qz00AqLn5NXxCxj/nDGsxlJqVfVxpmMbZTbH2XX+6PwOtoLXiH3wove0FMHSS1bDry9KeemygJm7zp
ahZfbazvaPJBT/+7Ckst0mTgE7QRItNcEsa7RfifdN+G8OP1vGQBG/z5x/ygCb9SKTgMnJZ/YCiho4UYTjiSrsum59qA3N2JXcET
HqQh2yJoh0whbypJtSzty1wOFew1ySalJy1i286lWwCe6RQND+77wUDL5YZuSCTJPcmuEXVFP3zSejb/p7FaGuigxSMOQ3CtBzCK
pEcReMipBor3t4yEneFRTp8CHIil+BRBW9RiJvjiajMdJs42IJ0kopOScH+f7/pd4IxCxhwagHR2aEIPxJI9UPNjgVfUjNR7fVy/
gF7vcn8LKV1rKqBAKtqasLKpnrorGVhX0gRIIDEyAyZR27m7Ymmq1NfjE5954gTRwnb7/3rm7wlouDQzRzXesDBIvRjAM5znEbDE
Pq32GJyhIzPSlJ2hlB2nK5SkHLGloKMYQVFbbn1UAZEUnPAkm75zLZkzfhuQqzjGiLZtY7T8ShcG7VYguWrL0hICMK05zoDgRK0W
tapefyo49H0sg2ekrcSBLyRRLqlwQYXnLbl2MRO1YWhfsEsYsa+GNhOOYRqOWoW1xy8RYuiHhhEEgof7/inkZroUu4qkyC1ZEKTp
TtnywX36gqwN2KI3w652a/XUiJ/tQehVS398wBEGiU2T2Ni4j8LN6yH2NdxLfwnmzuQpRydsotolDYUjhRAIQWuRH6emuAKPcXP7
svAhRWdatmdOlClY/N8dz6QyJkD+QV/QV4+Tm+S+iUvlERXFyc0nFvwOvBRH1gblj/uzC+9AdXS/IiAbR4n7KVkP/AeZgkED/fSE
K9uQRmw5YNcLMIMcTWrw8vQRQ5JZi09HOnYM+GzmIWLMbwcL8lU5poGFsaid+xrYw00e26j1AGZGQBhdSA5kzFx8TbcrfdCSllQ6
mpd0oUSDXq+Cax0TQA8OVWvLFkb+vjP39WRef6OBHb+NDdgi9OhBkijhhOu1BnqnppsPXB32OLiM4VTl3cy2j5CmP5I2LY6yNuHh
OUNbMMPQsXAQJC5pU4O1BCBKfKhHUSQS/VidgHBdjYFJFw3ueDuBoLVITxmdz31+ZAMccwOS3qOhOJ9wcToW4L03gFKK6FkuLDtW
dxXe4scQtDatZtWUbk7wjcICdkG3IIa25BAx4e/QPA9nUocxYoNher6hiSKNcqjW0yz8PVf77i5glVlMAYwd4dLCpiwFf7lTih+0
drD/GWxP37UYnyfepfqlZVhaPl/0rOpUyZ1ch0Z1CppOzDnYRx72JRIKepwRwOxFjRxqfZufF5Bx1rx7y43b2KdWrG7663H4+0Yr
r+x9n0CcNOIyzFFiDieAKEZF6/aEvBJhS5n2jwGfgPqUgfKnfmS7s/b1tE8YpZjuD6qiHjz13wR4ye3G8tove25BUG36RbtR3Niw
myv52PGg6XM9cY2IPpUAEkISAgFQuGo5VChIsasyl0xeyzYveHxyX38Aoi77BvinCBXAeEOQRQ5+D9pCB8ViHE1EB1j5/7czL80Y
AY6TPTcQIlGLHViRoQIARuYMV4tJY9rXqrzvwDhP9Rjhh3l6d/3/H3+/BU81US89v/RsYPtiavTlc2CXWpQ0oyt/lBBI7I0XC68y
G6NjmjF1wy6GojDXqA4Yhwa/uOKMPZwFa889NQF4NVtlkyQRw/9bSz74YYq/GQaQFzoBHdiK2fvifkP9/OEiZX0fyDkkyFBPO1oW
8qnsp3munFsaUIZV7e83rrtsLXt4QdF+e4EQvyYJj7i5hBScU4DGnwVffdICxEy18/nhGMVJm/QqdpuzrQh8s2f0/6Rd95xRppor
ec6wF8ebRJ7jyQukVwWteDTp0cfCTEB6BRpnPpfrfkeczGNEUuwoaL753ZLAKe881lgL5tPdVrSTsUtlwjsy4ZzjxlAHpZyQ8TQm
Azs5Pm2uWR3kWS7N9QOcKkMkKt4ZAGiIxsR0qSbP1ceZwV+YeyxWPkvgbi0GJYovB3redG1pWufFaQQGBh93p7qbQS7p68iAaYoe
ZkzMogdvz79ONIo00esf2KpCqsNkBjmtWZkDVe9yHazNXW5ah7/QDq028+D+eBGfHXcqtgLq6BGcpGi3yWo41O3D4qbuGnTF9FBR
/3d2VDa5DI9G4TnxB+ntthU2A0lVQDbKOTmuguKpHMsgnbkd/4K53xJ2rsfwv8SvP+tRg+WVPyIF7oOP/vV3LDAbp6eg7mUyiw3T
JE1icxWiz+u79cNvfb0nfPSYFBjZg3SL1kmx/qFzIdIxraZVHOACxVrg/m33uf2w9YL9Bjbfh/j6IBiMLfJN9a4AnjJEjIX5xw86
d6xc1OXMsdJhR/APXpmDyLIUvH+ENEQ1BJ+YvdJMhz8Zv6BWOKPta0QrIu3cL+9/a0Q6WzOVW7C78Ia7q8uu30vw1Uvl/rANqtZr
PaV+KksCiW+T7CYXzXKXDdQQGtwVuFGtSpV7dMrzjAI5WGi9hl++b/SPYt4z6/6T/97jl1AScTX9qTXGXt2Nkbt6BqWmPVvQ2/qb
yAj0BWneVTLYoJseEsMOXM5oPWE9obVgvdBkaa6R7og6BST/KkjMukv6rqHjnJtdLAXRuBfIcaBlql0Wun40w7XGIJ3JLwny4Y/a
/YVru6FWtUGej9IPxDabcUA1Q+Y8EcFTFYZglyMBDBUcKwjiAO/Z9p1g1Gw71OI8ALywemE+1/r71RdDpBD8KuMklDGAp9xZRyb/
ApDobO96GVsWAQ0ABVVWmSA35oU5q0skPUQuoUehlYHQ7MbaTN9+Noqdjc3UlJfE31OFzbEMyNkaFVWalrAlOOCDGaCTjkIdlwQy
b8B/qbUQQcaLmBqT6un4RBkrIpYAXyXXkjswAvxvIsFQvrM70OjEfebeNpi2HdeoN+E4OJk6dPL5yqoKpUJVivrqrgwKagXjbqKd
sWGuCLTMZsR9MoUZJCOfBB7OHXR6Lw1M34ncFn5XNpFQTwQKxEVYqjIY3wxub1lAvEy4z6gjgJN3NBwYZREYG0WZ3uMGHlIfQ0Ia
lU8zkSMNII78GM633c8t3rO8M9AgLSjQQ3g9CueLUpLk50jBGH/CkfelhkZBuyRSJG/ZPsRvaajeDoZyPU16uve7dzpgRy7f1I1B
4SeT1nZWtcV1F3v7YdTQO4aiFCcdtrEejVlgm17NUa7r85/WNb43ryEtNiwnkoJ9DxKtlKmtBRQsKyB4TfNFGKHEbvR6WHbonFWv
MCuUtwXgK7OxUZXVLjaZLBR7WR5wy8RUo0oe5VGmm7fhw/enyNwh+m/5fVEIJTFauEcGOXBr9Ms/A+Ysw6Y/CkSHqWA9Gp6u28Us
AtdSPxG8gKHDq93AY/BRjBiawG9WaJBoqGGk+gbJ/Rkn1adRqDPJVXdFiBnTU7EYY3eqd4gC4kzs78Ixpw8Z4uj3C+613NXJd5J8
TM1CwAhjfIgL+BPZnoHQUzBczcNjjyJFK4OR0XTmnI/iddNesP07CwZBGCKhfkJIw/az0/0rfEwythhef+Ztb2RIzFjkNT8ZmCx5
bdwQT4uMgbEOQu9O/GGbGu6G9gBX9FkzIT+lmRsuT/JMkznUwnBCqabQ7hYH7UNKO7UgYFHdzhapEC3gsEJjUyawSlj+oe4ERp95
yKpLRoxDBHy8Ln7tstXKD+EP1AN61WmvdvBGbrT2Lybd7Wt5ZgbAbcMKOHxb7RSGGt1JLEKgkF9+M5EKjNN4254bStWvMsv6idi6
I2jGYgDCMx1nicEwYOvveG3VaTuNN3W9WFDb8CLD3WnzTFOy1w5BirsSKT/GNPGBa0wzwlrR+L8wbH+Z9tsx8kwgvkOE7iR1s2+K
FflNaaK1rLZp4YTJ27T1hPInfWy+Rr1OumJaWP/FxMjgLeAjOWXXOnaAlaSmrXvh7z6RjpwrSsCz879qs0DFVJYO1MrR7dRmaPrF
MDuhDI4D1ads7fvIow2cLA8tS9bNpmnOCI5I/iVf6dqMqe4sJBDSd47ofJXA8T1nbRKncNPFDljPUvowdqIa5rR8I60pKHFBNX5W
H2E/mbXZwB6VRoDh4JUsO0RBu51ec9PfDVIWfiyhtFwPKvkAQx+4TAL+L1osoUszJCRl2Cu275z74gbb5oIcBh9W1XLF0C5AHyuu
tz36pv1kSywozSg940u3g7U0G0iGvauogThp6V71FbZpU+MwMV6klxVmk/1694fyE+BQ/akiMFZKbkEgZwHkk2Mr+rX8AJxyZ/M6
TPVUxVP3rEs5YycBi+Z0jgKHvQy4o7EO/RkJQLZ5/uYnvQPQezXgPAawiw+XyivQetNmhzfXjQUj9g0Kco4AaCE9T5ePlio42yRn
0v0YMvIj6RMrTzcbmmtLGVaXG3zmRRlGdEBLgsji/5P1sqwLp11adkapQelt9p5VXiZt7NgJ8dzxaqMGXO9/HQWXfQcHn9Qek8ZO
vKyReLaVF1GddkUmeCo/KbH7hMGZ9GWoWurt37+UtELADNNvvVHE8OlDbJZjmhb7Gomtwf/f6evL8qyndA6okjA0/slIBgFa1UyQ
pRBaO0PSrnR7fbcoCXUKBhk+fkRGmw0D1v1lb5K8vykidZ+UggMZQk9IFny87aCbuLiYP+ISlD3bscvAKvvAO+bqTo5UEUhCapo7
az6p9COK+QsjE+Dm8QzoOwLcngJC0tt57+RUA3Yn3K+XlkhE3TArkRVCD2ho41mAiqmHhx3u9dwiG88bztxeNig/nubo8XAxqa9W
Rn/mmSSqn96ENX9yyXvT6QJKD4xO0752F+LSPTcXqv1Gf7FbddSAQxXs6s6EVDpwam6nuY3QROAPa5maduUB7y3eFWaefvPkzjTg
mIARQtnWNy75ZTaAC5W7FThWsQ3jzfECW3l994pP0+StmrBQd7+a0y/o2F4isiqmytx6vDOYrydTl4tiD9R1KZ6xhMFE4qyb/oPx
kxny/svD7NOjXApLRbUtk95/M0w28CvVLMqB5C8r5vIJDzozB2VjxfjZm5/CDWJ1xeFJOJwZIr5T0yyfR8VqX1U6q1+u4xJ8jCaH
KkzdOGHxR2CLw4T/nREXkltFr5DBjvMzICui4+NiNtcyN54wb7x1XOkZasSuHRjn3W4Trg7qbF6HngoBF1Vu0ghk1GTOldBuRjca
44PtbIGXIZ8RInb4MKLf0MpdRnrnwdzSAf9aIVSP/wEL10S3jubre9Q6mVY9hy1J6YANe/5ArKMsX6i0H/In11C0bDoxSghJ7j/3
z5l0ZiIOwknU+bGMRK0Okjig2ISF0yMY5WFUilgHY9sHYmUy1wmXjkNHpEzhXrHkwBgELGrZT2YdiicdlOT+dIEXJJzJV8n2EGZ0
JDvrJ/RTEv7Scynvbmq44dLdmuzho/XnEvtnXRT54i7ZtzBwB99U9PzVCrtQD1OBTmiUjPROrbOTxPq5gnhHp5OtsysLF8S74txm
H8CHsl41YYgJP8mYM/bGvyLMZZptfGYxaznPLaNbzOVa54Nv52et7stsufxeit4jD83kZQQbn9CC60nD5KPW3tkQxAKI3IyyYdVM
3kKTMks8qZ1gWj6JyT428ZiuCfl+ZV8VpxA7NdS+mVVNyE+6cLZrPIVRcrxBgL+zyDxKHouE3+rvM/zTdFu9pnLcF2sHZYQSeIUF
JaxOQD+wAMC+hUsBzQ3P1gPkGeMQFBKEGE+6DqfOhSFLHT05ba8JOjDY+NM1YIwXCGFl0W1+Ze0BBZdP9owNTrs4yijS/XWj0HES
OB7Jr1syHDjkbWTrGOZMYN4v+Bs3YGIYUuaJUhFeRGApsoV+rM+C/XzGR/gt7qfYyopde4vTNIvehv2nD3WFggo4BBISAWnW6Vok
jsW1CtNTSkBgfgk5xRambOXTX1RDfBaK3ZAI4RjtcUVGlCHQwdgkayfK6Xt6t12iwBtBM16nsD3Aye6E2+Bfuj8QRhj6qfCQJ+mZ
Y3ZMK9DChLhEdmzfa2dt8ltME00uT/na8abZ/WVYYsVHO8DQWDTFTQcLbhhXP3FFKeQpwiNgBv1WUu2dqTEjpyuOVaB68z1u6hIZ
iJ83UA6VNSGmuZDELG/ZY+pG5yMS5UhE1W29HyHsbpkqrhB5LvYwAUOQPjWKFE8hoF41pFn+ivCWjPX9Pu1SUgaN9vhgfWXX57rv
3QlbcpoAkDkOVhbsEfhZJtqM4hhQGvJtWrMAqzmMJch9WUeI9Vl/8iuA8y8FHCgPSIw16Ah7A1LiaA+ItuEw9jYYOSVqvesymSBH
qJEmZ+9cdwFShR8GAq2TOyuZQbYOHSENXkHNDdotR+WYeUF4aY/iOa132d480qokXtO2BHXh3NsaMIvicpQ7Bcg5TY6afF4/W1nd
saCHdfms45TqIF7kh/qQLhbzyzyMiIckMtbVZahHFZAj+yHIm5W7TnPsiTCU+QwSF7mVQzJgKWpchdzpb2SFFGCFWeIp0cmeE9Sl
CBj9xkFYtDkvumSnNZnGAImQ7UNV8kViMQAS+DmRpH4cTT1qAf2as7b0SB/WGdZVFnAP1FqUmt/BLmmW3zpvYcaZ9TSpd+LpV07m
UyMaYFkMx2Gr32M8J1vhFvuj40ODFAdvUkP5mC+JZPJPMXdcunSwzLKVpp39jQeg7ZzQgp0yVq13GrKXN5K+39xCRlI4QcrT7czU
6eHbVgf6Aj5jl+UPNmZ0kUH+TUQPb/QYZa8PaQrPdJn8x68+W6hwrgLoFRQUj939OKhDZpsWsNoROp4G5a4iHSDPqqVSwf1SBwdp
/M/+rcykCZH4ODhcL2aFBNEZY4Ka9QvvNUcpO2byUcZbW+BlnHgXCC0R43r9KJRZv1TnWxfwWK31VUgaQrXeSpm59pdLPGaCmUsW
3LL/S5lqR4JfPG0VkpNbVJ39xO6JwCG8KAG+SMwoL0OTBgbmzlLFTpu3osG+/Axwr3gq4qxkafb5NBuLfYquPSH/m0yxaVx1Y4Cq
USCfg8MPoRgB3ZeKy4+zkNJJq5KoCB5TRiW9DCB07pe7tCwlTFBJhutKGh8xPqZ9zeTidCvkETJjE2tqUqN2u1ynJ3kf/dqZ25uH
4gsREMfCSObUkOPz4Gl11P/YXENYM31e0rt29XmajEcy7E580zh9YB7ri/Wr3UuBmoZMdsyNEpCvcO4eFHJA0qCNqPHUk3HtVCG4
tsh+C6u2B6tz+oFPOZagQg7DQaesDj246+cEVXNWykto59WBcbuHmlowv3OrtgMR/G2+os007NLLn9a9RNKWJkoz7W94Ki2p4665
75MfY8Pn8zVjV/l2utOaFLBmjtNDAYWuvord5cQ2XnwA3EmHIeKP4XIsoiBomnh6w6ZsKdtX7Hl8Ibf7u67o/2T+eMunctCf3BzB
Mn/S6nsGO8V+hxWaql0EYkKHHjsllkvUPHxTOn2SX9QDL8ERvCMzENJPUwxtbKErb53fVmgyn4qGVhnzgv+7MKOMJ1ao0aN6pn2Q
TkmKa8BEzu8SPpqTFtiBlPI0Cu6RT1R3tQe8V9Pu+n3f2Y2ICJzS1rVGpmIHh3z3R2l+tDfds6MakoCcpTsN8HuoPWeNvu9ezT3Z
h9wSEBNwxP5BV/Z9M7iRjGwV/Q5b7mO135L7+08kRO76nQqjzm0XgFAtDGaM+JYd04Pje/hrGvkbv1zbUkUXBD//fKMM5GR3SjyR
loj2QvSOYnaRUDnQc3wopLsrJ8bKDCFZoeSLDW8SSl+EjUfMNA6Pmb78egVuBZTdrd5ahxYdt4Aypn64Z/9vQAWZTBpo5QKuRFma
sZQrTc3zY0Cm7CntXHdLCFu68aIJfOwWQQl7hDZ7hIwmn3ERBSpthj4/hXQ8Zo5NpJoAnRbsTHjkIZZM2BD5L7udlesWXK98EsFd
nXF5clHsiAVJW+11wSp709GnRY6+WwGhX0Jure5Ebpy2f0EmebYCm4Yrbs34V/1Z1obmokv/odZVK59JQZGnkDsiPwB1FG9K1XrO
kaDAwPL4czBcVzCpPPzGXjmPBnm0IUmGB3R1SKQyobbWXrvDGveERaBZu2x3fi/pj1wJPdt0RSt63ooJ1oaIiCPcHEdlNxnu2Zc0
gO+vW+w4PIGer1KW8Gn1f4yajwRx+NqY24EiAO8jWufq9gU9CujoqRcykTV24DNlukxrLsDIjO3A1uD4V1VZ2d3+MAScevJlk4lX
98b/TP2eDoOAtxArK3uU3h1bql5ppa87h1sZ2abqphRbxDK/5XYKIGikBT1TPtUc2u2OnygzwRkAq8IwAZJWEO11rqF+ol+7s45Y
rr9dZ1M4h6cOmVS0+/XGoxU9qlgUCETkyQ/8TGKVcKnnVjf6F9xEQoPZvEnAJqaq5Gcw5aC2438Riiv/Dt8Ar+sRdYSbnVPrd6Tv
K4YjWEA5/gazMKDY71EvMEUSpqcleyh/8VGAl9WWebofKAUao3jeBpWjRRoO7nZauKGpdqJqxIevrCWR2TDknDUypsYtV0IWIol+
VpCKix0blFE+UNKbnZHfypQEsAOWLjHstoswzavejeS7kd+nFcvbvGxrayMWItsBumV8d7gCx8rAiS3CchnD+xfuBk18cISR42E1
JFcNS50nvqyxwP4ZoEoBNqMUWHj+h/UxgAibB/yo9LiuV/ONSm8dN6yGZ7ClJwMVw9hgYasaX6je9NjgbuUVispGsAZl1ttzZZOc
AwuCiTdgiPNqpzseFM4OEnbBcHIgSZ3/HoboeLw/eK5sTf/rDeoZWImPMVt3akDhBeqJyMg3eMsZVRNVtg6VrMszS2dm/0jm8zzT
Ze9yKJE9gTJLFf5gzcck45cGje+gvksPLERmnxWMKAmUNTpjuC/oc7GgWKvtjuTUCfnWTgplS+LWaas85LIFhiR4FH8gUNFSkEXX
ETAkxRNt1I3IqCaUO9oPNz55/8/4AA5DpWjsUFPQJhYUBGJdDRnnVcva/M9G71mL75QP86lgyc09fGR86zPZNoOfF5JGWacwSb5x
UQEq0qI9cchkWAAzEyUeAoglNWmdUrAZE3/AW2UivPwAaHDdJvysk0Wmz1COobSZr2CEP7It7X0ziANrmIGJfOzzEDmaCumVTOWT
OylS6xf0JJrY9JRgv7Ij+hBEPZqzWqNvU9Z4wnPNBB0fbsOGno+LH1z9aLydfkxeH0u1FDVQSul4UVnc87J1k+XUVPsRH1FFE3yK
qzwq4tXgi58UtcdPMr5gFqGEvsUimrU9XwmC0YuRzTq2BlDPoPNVM3XnXIcNJUoLPGOuh9MF0RERa6GGLaCjLE9uvJnGrSrWFw32
07KD8CtgOfvRaLXpF4OHcC4C0tCpGOLwu9E6CGNhUz6j7BIhyFVU82QhSksJl8Xfyfhrl0WnSEuHLnlzVIp2gn2vPivfgBir1/7S
ZkVTmvdObG4uM6ramj7IgWDpJR2cMG5+YO3DvCCWnz16IGoujgOn5UbXcB+pgkDgxFr78ntyVATngPhVdhyLwvrHANAtzCnJ+3Q2
O9ExNc2UgMb+B6psc3z03U6ACOBM+M1PD114lBFYmemkeU4yFL1ly4Xj7GSGFSSic8cB01AyUZWvT4N/4Z3tIs/G4OGq8zJvj9Vl
9wtGxPCTy49NoUIZtOD/YLwOKqStjBuTBh21UBr9C3p0t4pudnBm2gjrBYm8decxZkTECPNPnCxd0BjFzuMQergT1+ZHCNcMc6pU
/Y6NKKZYwJ1UT/QO7DLeKd0u5BYxGZD1AvQCj/UYyP6N1LCLGWHFVmKGfTVT3UxQQUD8HlXsJsekC+fqBduZffAnfmnR8t4pYqhl
Qi1+l0TUx4j/ysOv4jbGR65yjiKZX1khHZ0i89klR/ed20xbzb7YrYiQ/gtmFej+ExEb+oyBu6fFINJ/3f8aaGqBaWv9N7F5ZhPm
ryR5T+lfqpa9CVXzWuStzoDQRl+Eg2vfQQcqxSBeks+DbzOa7Eq10mFdKOq40TUWmgtK/fGTxtY9vP9k7JdV9R+jofd6c732Lr9f
tVBdGvSnrEt6cMhu5FlWHCABVDh3lzRcdcb2vtsXyMRv1K2mQtljH+dI3fsTww+mstLivgr3btgjto0NZqoRXWxlQjZNQtFk45Uh
R5Ht4zSPdEs8UI4r/NabRV0iLA4dUROoWENwMTvH9cT3WKUNHaBbVlemLn04+MpuX6dOuK1YTKxqyQj+S711RGlDWRpQ63VkkTso
S2kyesIbrixwICGkmOzlD98AQwDYvRpUDiruFvmzTMKYibOsqgNpTXCWITFuh/EALUDO2hMby5N2bbZlqT81EvBw6TNA0IinuA5c
IHL1qSxJcDVP96lX3N8tbvSvAjAFitiqoIxOG7OpM8mplYFaWxXTf4olwPSNBMCFaVPtdoAS1vFUQTzoorVjdS/SgnWuDHMXPBGD
/5IqPLD7EkR24LoLLJ0n6RyY+RKzjx0kPDUF9gMb+XC9bhwoL+Er97cBQrMA32K1l4/4VcikB5rshK93xLxa06v8AJPEVl4GNeMM
eSqajnGQu042PgPMGH7Iq6zFVUZ0mwRNuuAPZcvSMsHIEEsM9yJ7WolHHU9pqKerWkVuo3CnrwGzklWU80o+FnsaSwzgUYvAswtf
6qYdBhscy7SwvTltpKwKuqnfSeJcpuIPvnJSvBjeFQORJVdRHaWP0Azit61QhZetwAhe1dwDvh4Eu5HS6LSY+7c6wTh5XABPvGJH
AFrmHHE5rtdT5jv1ECEmUzm5PhyBJ2NHzwqeDVT3OhBdlEmB4LuDaW9UYpAB2l9qXhdS7lei3ADaCQ++KmmWobYr9teADba3HhWK
oUuL30UhGmCALzkFkB7w/zv5v7FFNEqAak2m/M1u+XvBPuHTL8bNOP0lDDlG40dp97aBMZWDM29lo+gjHSyQIxWlwhSoXvD9FQec
wCZI+4UE5506OgNo9Z0ZbcNnhN01Wu7qSRVpp6pzA2b70HOel3wVSCApdOYtW0/lwizhxY2FPlm1IFd8nefIIZGOf+NQJqc5C6tV
YusyFbGF0eQCngB7n+zPN8YLX7+NEFJqfLL6TODD8QeF1zmaGWqsbpkNexf+wiO/CLTSHhYYgFjkyxlaIG2b0EtnvsmdsCCfXSJM
h5qjCziiNzTYc1BC3/5HuCvTMnU6C45O2Re8735ZqBoR2jkntnFjtqsaKZ0txyemt8IkrLY3M8PAfO7R3ITaB1u2T8GkEhEZfpU5
fLYtFHI99RSXoUfMCYs6CbywFkLakC5mfqqcBwi2fn1Z+8eVnAT37rKbmwzQiDR4uuZAS9ULppoOnC6D2F+EpFxNIM8QjohSnpI+
8vwJgj2GnWr+COnP0yt2yxKVFPx98TF64UToD5x3kV8fmbcOUKNuEUrBUaNVr8HuO7jhu5HI3TOzYN8nN0jxADLgpmqTUul4owRA
urlYQyYmnZ3x76b1dhBVGO8WVTD5wMe0plLWLJikUPymDfwSVePR/DvqppJJVv3FpdYSL1xYLsLMNYYUrsKjakGQLvri2pLmvdRF
UH1QuYTKyzhNgjMXg09Jlsho0NQ+298KBQ0gPYi4tMZJosMGOTpNsl2kYDWseXx4sL6Zp8uno/2T3hV85hgIsonzNZt7OD+dnFbV
IbMwdBzT3doyP878tXLeVdVZTlMMI5uk2w1oy19E4rTpgjdNOtF+tWQJnSM5GY8GJ8pEVQVWH11jzk6iienCXAqK62TqBFFPrb/U
KlanmTACOxDTN3snICOl5Q2L/LWWLPYV9q/gfmFgjpXxFQuu89Is6qBHIBL5nnakYARt+GOghe0xwBhkoSnrXCiXaFzk29FLIySU
Y9t9ljgPqDhKO/ehTKkLKGa85oquMGanTRkOvqOVxRw4hOF4glRuAJRnitm+poGDYmXUPK+E5wUanOV0l009qTJaBkIAYIJjuYiR
mJc4zgf/jpiBCHOxJi6qjVJIp8MfuxJz8r+MEZY6GsakTtE4wI2cqeyBoKX2CiN1pjojvgr6KrNi9gVPSAlFysTeYCcT7wNquwL6
FIrQYqIJe/GcwqVTEHab6uTdQaPQSoq/u6H5DIyuBmveCda7seXVeOPVzh6BLLsj1D+ZPr/YnB2BV4BToSFW2nLXYdbQlNSmiFE8
cIc0zFI231qf7A7CN1xS4u1KcVWW8W5qLjkCsf1LrULMjgOaWsdLe5K/5vYkhtYvlJvg5JhqtZSPN2bQ65GHx3VUqhIo8RiCNv9B
zePTRd6JsMWWtEnKixW/Qzuz3LkQCxjydnevX3zna7TsBkmLXgHKxjeJXOut5KP7buXjN9iBModkeRNM3I47pDQXahisFuayiq+k
9U7meiNEEQqEivviLZKegvPECd+deIswH0YTr8fBe1VnbX2Hn7DDtD4ZG3jfW1U/duNPq90yxYeTcAcJVhCQ757hJ+D7BFZo43dT
K2+lCixX8lxMt5MYdGIMMmB+QqceJony+a/o2fbFnzGcAqDtUlIElZDB1gNGP8rKNGBvKFe9CQq0DajtmQ6CUgDxiJriCA1EvtjN
q03lA4yUKptnegZMZSNiJ+aX4AVWCcL9iiJf5NoB9zfX5YRjfhNqxe9ectVHbpCHcZl60yNvBTjwNrjpwF6bXn94JumLgadXYNUP
+hbL711+Y+ELrAL49LD4bXb2HEORpYiVRGPVLrq6l65ZVgNbFRpcxoSUzwI3Ja3ScM5POvDAdC8ngIrEYN+K+4AYLbXUOCiqED6L
MBgXgIIN7FU++VH8JTuC7A7j+oBHilhv1b6FXOkjRp3VKc7JZTshdf27Hu20huLca5wBq+fdknFUrij5R5KM7311CaS4r3nHpq0D
UlbAGBCCB702RkhzdLCUS5RlqhCboJt0zT+WaoahZOxaTb4P4kotXTddmggs+K8aoH9a5bWU3wLKhOearMQdvKlsG1pYCfRmQgrM
tTwki3PP84vCWxw4HG1jyvsYFlrBRIacKxsRbFUPfpkUTIjGH6/8UFkKx3lpnSvn3fpQW/fr84cRueuSNgfL54wgGe06wS/F5aSL
/9RRveQWCIkqMQG9awsHM2U6w+BGSpK2td3sJjVyW16UyY6KzCtZjc4R7tzbhK5UFWWsKUvIczCKhPYt2J3g0DMJemIMjobmg5bD
VGRHnWzDtGi4JlZtradT3anRH/djZqDE3oMHVvxZKW2oC/EzmoUBuN9VZJfsA1oPgzIYBFWJe7UPUdzenW0jgkh+pC31zYF7yEzR
RJOP23LM/Ur+ij5HVyTWJPW0XUZ+fqymLdTRqZ8EZ70S1DoVWB7hXcQ9OdeMbZ8u7GFd/db0C1TLZszXC5HiS8xikBEUnUPHZm+v
QnuSvVywX9O2CjVioO2OjKd3Hv8pE25fE4GsmhVTc7x49SizVbPw6Ei/mCpBPJfMEVn5WJirgVEO5QHH2HaSzfKIry+cV3Wc9EfM
BqDgW/TA84SLSv46A5G+u3GkLSoxpa7QCjNCrInKdo5Npj9v274FaFHV355VUHvVlqges2czbm8RCe7er/SEEsgcUkiIULBieOTZ
pG2pI9RsQQzeXVoLQXBI2fIMUo9VSyBeHeKsFX2Rvw48WZx97+KHBG01Xl78HQTzCmQnrAClx+lnAUR9gfCXyRefWU5GBBRph7Fs
JKRkG7oA9LqVlPQ2a5hkkFP4wR6dcBMIIe7LeBnFYRwSv8OWqYzqVm14kcyDeT15qNos+L9o2oWoXhNTFxXgeL0EFyxKNYnI2JmE
hIxZmJSGQExZAyO6FNdlLulPnBCTK2xS5wCo7z/tcKVzQkbmwUMskB/wL7MrJmhsn+oV971e2G+OU7ds353XM7fCGJ57KTWwkb1R
/jAQPSktc8jDtmA9xuDY39P1mTxOI0aKJJS4+WNbFQRFkxCl/xUI9QPDbbr0+Pf6eM6SeHJtYMw0LIdwBXGwqtGBKMQvvcB4rV/z
2KwHPdDpGDYEReKVp/9cBbBln/qxQpFedlNfRdSlXToBdQBgHly0FogHeCNPoXv1bGoxYLbvJspgKF08pu1AnAuBnVA56WnLECdl
gbk2bNolmDzmaYQHgOpq4YMiwMSetu0yg28C6olTUil5NLx5ApoLBehjf1oA2q98WIsDbJa54u8SxMXJsUn7k72NxQvUHeIwFMcy
EAV8ZaLI9QfZ2DYlyOZDv8azRPbo4GJ+7yYZyAF2DS5Sf3LANiQ4r5yMcvhvqctArVcAv+93jBB940VoDHs9vAuNZCJBUdYrFgH7
8bmMFKwG37wqgEL8RggtpJnbmRDJpHSNFUdoSlXhRYm1YB0gv5R0yCBFnxbbgxJIuKUPL1M6BupOU9OgLWIsyiDuc+XVwiJwI6bO
4T3eTfjk0BDQn2CHKNmuymBx/aMPbViMQJ99Dr4niBCTdQMQXSTvA6sxMD9YxUplpjlnRCzyiyGwnvwSnhSUcOh9PacehCQ7XD4y
052s6KexDhDDU/xoQ8qifpWD/Fmnsm6nNFsocuAbgdX835NGyaCa9GuY2NESCDEUZv8CSOD7eCqa73heM/zA9VtudqVKHSBNT/y/
qNqff4+0Fm81AO9jIiZpvINmfaPkWHzdEZzvRW3bVD6yDyTkVX6t7XshkD0sj0ah4XuurYmJGrUaPBWrD6fqxhslX2BgM3BftnBf
zuBGgyu0Ttc1iVE8SI6HWNr07SQYD99l78BUvx14lgdgJ6DboCUGfLSuDYiZcEm5uukyEIjTtEsWx7UWN60nzyoxqGEXVvZtUPKQ
n7DAcwtFo34/Xvjh+whtgZ+G4Kw7JLC9rdR93jJc68r6xPQxxM0oSgspEY+ypG3+JHDcBBkVvge0S6JIqJRVvqdo9fJOv2JzweqK
2psHz97EqhuUB/PetIkeifOcZgylHvMpiSc+lRuzWzvQfTruWKiVm9XqmxC7lNm793iM9VAMB4TO0xNOdXlcNAKzcllmXDDk7AjQ
lv6Cn2m4DubMJQJqHdkEkd03fBLhcOkM8xrbJ3r/Y3/G7/GiVO3AWZh0wx4vHlEwWlx5bY3H4IvP7Gcti6Ems6CB8oAmtvSXdMDU
jqcp2DTKIZecoZICh9P74elHUMpUx64VKl4b4TDEotVp1OyISee8DbUDAsliRGoecsAZp2+mOeI0WNI2aHpElvJ+MbsvtVlkOH4S
kxM9ij8xujIwN649ltJVjI5iqKHxw6glerQQ/NCugfKhzoD2e2MBYvw9WsY/H9fV+03h1TZ+xD3U5TVJ3Mi6ZLP2uUlGouhlJkMq
feTWz5oTpqWcWnESR4FZX71y+EZKFtnz+dzogoEReLzF8PnjkfakhG4Q96Ca0huk1hZZhMTqfxrqu9y/owi3HSZqXCZUNx0LKpXZ
WDf3wMi8AF/JLUrUf29E1wZXBh4W8XADdPsCILEhS3hzJQAevnz5KwgFx5/YCsrI/bkCpuaxjLcq9M9/ZTnAxm9ZNeeLxDkL9bu6
u+Lux/Pv56yz0FsgEqDKS+dkjBUgsN+D+yaRTCiaqqrNKrUOH5z9AS58ELrlHbo8fEdbUu12TGJFXKkWMR8Z8dZtmM+vZVMDLwsJ
G5hyz68e1zHd2c69VU5vNcKxUyQVrOUqgqPl0gYbxXOO2+3BC6w2kG8jf/XwiLdBxS6jbbgagDcA9yI0Dt4owdWTqLIynINN2mmb
pGQisgr74iaYM53py0iFPBJxe9SEo6GP6/KeQyLSxMNDaewmyn7cob9CaC3Bxf9d6qZbNrqSYyruhioH7CB7xoMoq4V85iokBUUW
ymGzfp0T/QB9hkb26S9fC+KhIlQeBbX2VF86/qQ/d9GqxdwpCutZsQ9Xs7v+F3WNfRGmJYMlFkixDCgSknmJGsm0/blLt9k5NIgw
rz+P2stdAXvyPzxnh1J/gfipOucnhxGnKxlA91YbSblKP1Ja0zUba8an90dtGUJ/Lyf6znuJNpB1JdUGdcAkKmpNL7Cv3CckV6b7
981Zb+YKjlJNkn9ZKVX6092PusViYwHbnPSS2b8WnDNAxJ+oHQzTmwf3SZXSDPyb91/SNQrSJtGjOgzAPTceZsgTxVxgMCLdYAPo
SVSCQzO71ed3DrnaKlQyOlrHFGM8vruItrXJBmmgP+kfozAdcZajYPyeOHZPJNoEDkdvZz2fXwbP6xAdqKChWpp+HH8Zk5Zr1INB
6GdGEvOZaqiompNJCmh0J8ayFGEKIR2lAnG7ib+nlhrXnaz1/7gWl4y/jtFtUzKbgRRQAk0aF8tXkyb5pfS72C9+P9glQ5h9ixr5
jg+hm2hsiVwSoURREMqupZ5JJoP/ezbwVlIj7c/wGDUg+wjwjCk6RshQ0CyKF+/OwUFLhnbP9ozbXRfzEh+w/P5SUHkpdI9fo9/J
kHvkIueC/2o0WABuW2E4tXuJzzzuy78aqebAb/pt66U/tNauvMlbBk7H9yeaOFn16gwUWsg+4WEiXoDHM9QTgaKS+k9sX9pFexZk
0YFKlGbipC8P1DmT0v5kbFO7JsgnAxsERWqafzLkdM5plRrRTLTkfDNF6eCawdvjE1eGuowBv0JUn9Yve0v79a2AFlKbpWJasaVJ
uTC3k0vOcJf+yBm71LX6/rWyBsatjsbEOFksIRU3lykl7uhWGIMYT8oQKg/ZDgNJrssXsr347ntHFY71lt6JgVM+s6ujCG2g4XRd
2ihPWpnWNpmAfLaDd69eAvRZq5tgc8ZYSMuVT4dF/FG8oBF6IzQ6cEOsQ0RyynIPM0YBovsv3A/jpH4Lrlpg7nsQN7uxvyFHoprU
xRw4Xu1EoeDs2mdyLYzFJQe8yp4EcosArwk/90lA096/EVzNDaOI/GzDFFPEtK5esJ5kYxmMZcrI1+zCF3ygG+r0GsknuIdiNSA+
E9+mMbK+5sgtE59PIbRxB8jAHWh9klkDHgHKwp52Mwooa+1+IQM2ls0lAcCiOUQFaBu7HtD1D5/yHI5w52cN5+djVX56FuvZF6Ip
+pDmDQ9zYsl5JsE1pn9jYv6/jH2g0+Bmt+e2xN7jtN81fAqimBd1fiLU4ohoowwFZlHhSekPSsqiPtKcO6lFlAUVdotJsPtewz/T
esQ1uBJ2BkcB3JSq00t1N698dV++p0RyTAF/oYObhkDevmZH3m+SfA9GR+BRjxsX3nB6LMcs2m81+dewUEB/lK2fcXyGp5f89zgI
ufc6NmuwnnRcqeqCfGRJ56zSIoK1/42gptWlT0RDV9KEW4HauFl5c/6wAvgpE53w+ofY48E7yjx0lLsw9BV8SDMCzgQK9jh1+B+n
HZB+TCgF7G1E2mmVlHJFXAy68GOkMZhQ7JyomojQ1MSEWtu+mESoi87EjeYhY/vj02M+De2haZ/e2g+n1hqPvvqHhSOqZwRqVQ9D
MHF5dVEzFdiDbDBSmLteVVXfojBROKj6u/IO63Ayy9jRaUfac/RYvPv5PMdyheRKTH66q/i+Nq1Bz+mAEQaPLAdGSPcpAU07Yti7
cBQxFAi2AgyRxCsTnZSyZVq+DJ353KU3KTsFmyNkTrgsv/2gQMS1WHyhKzvXwrcqNASdRV3LfPi0aysarsREUaTQNKbPtMBdyXD6
eFtnT+et9SM/iItU6P/TaJzZaXxW8sQYNAP/GRXYyD4uKKJ42iFKNeFJ1dwFYvB28p3xV74s8PutS2D5Ck4FnrKubDEYar/+8TFA
K/yHT8uOcW3qxA7AiwZXa4erSPxhiiRQM79wUN1IEhVX79REHyn0vcoC2j4MTN6Zwm972t1KTo2T6Zb6DDrm8Yqexg4E/4YqrM0L
aoLaO3Nr1iERXDDKCFpBmySqPACLSgkZ/CDxvsfsYDC2WomRjDHIgqD4NG5BMb08CtVTPefQ8r6PhAH2NYqyiYuBSTgwC4QAgGRw
j6m4a7kPImGjsCth2ef70qbv+quFTDhBvpkNf6rqyLbzFsB3hGNxVK0dwJBawnx/wSjR9HstcsEI6b5zDi6YWRRht3Rx0XGhXgr9
r720dpEM6NNL9C4BCTNHbj5yMCr+VMF9Bls0ZGGlXv7kyr/3Dcqh5HBqiogAxbyUCnXVWuf6F+i3ly1kZ1Smw+7iPS8pHrru5ijv
L9AP9nbr/CwQl2ysdjmLiHMa6QpdrB4qy05+Iwr6vjp+AUKILYlVNdLUgBIWb4hdEftX6hzzpo3p+6yOM0xUmoTa3SCjmcLxN2FT
yxrhMTNbsaPoyk6mcNstKGkXncbLUjiOLw4OvOkvKXz8GKTVbd6U6WTc81dgpiiYIq1bCYpds8Ut9UlDyOavPkIqgm2MFxpsvfgH
2HbwKZ8oAXHYnFaxOnlZys/aB5AQLyGwaJzk6gbPuBTEv5JpLMjzLVAtH5VWIfChtSW8NFuY1McERaU2r1psvsF/ttoF+Tl3OysW
TUDEa2CHpmAuV6d1iBuy9bobk7Jv3rlO/uAZ2uQM+AZK8Uv6mmX5kQFoOSRSxZrdn3iiS1LJW8dF4jqCS79dYYeHOzfahb0LpF0P
wUQfRF2B4mFqWHTsP+MgO20N/MEIdCO3H/ln6SIbxlXZDa1bkYFlo9AqY62LrXokv6WXpHepltk6ABAqcwLgiEBiZPHVndiwof3n
PWVE69TjiwguPpCX7EEeghmjbNaAwOH4X+HX2hCFRd7u5fqcgp0mywAjU2kEABc1PfVzXPDLNTFcs3bt8FnwPRdCAVDvabMN8Trc
mBC00QKKuV7DExLNtK0Wq7QePuCGgiiyI8afd0GO0Bu5Y/1oxOIMDd/Law/xJZuqrokuV4DtQxbPYKwxgmkvKy9YKZ9tZYNcIKhM
W9Z4L1L3zabkpzf4/IbamQ/h0X5xnCtFLUOWcelv2+0yJzlG+LnCt+nLm5Y/HTO9wE+xzcDGu95O/8+O4sTCuCatA46jim6nUxyl
MtLiJXFvtYChC4pyGfOU4Oh32tEF83T3zl+TwAk+40KLqHBefhvvSGvnPGB9ZX6S7+RMlyg7cZ3Otl9xUhimYBMNEkp1XRwczJIw
bK/XmL14Zz6wPXQdxobOHkEduO53bnpQsz5z4W2V4UH8cYPse03VF6oquyoxZjdUPztPjhBUUkLlcpc17bxEnESlSGkeyEwOp9ZR
Wp21kQmLKkYMA2+MUeZMRhof6WtbWY8IEY3+5RiC1HfJs1uWuaLfFjmdslelZ4C1FDCvbOVCIFPYVjRlkmbUuZ6oJz8nSttv0a7Q
p/iQNXYxr+Qj88/pEDcshlFdjdMIR1f8bHuCqoXMEFzJiMqTXnw7pBrvDK1pOSyaaIGcJQT93vmELokC5u7owTF/26RqgiUdk2mB
UiQTdHpBsmoeq+EXXA4sQY+OmP4RqgeI7akWAaXghypGe4SdtudnhBcb0+2tkmJzx+Rv92OoksNglLiLi8KGjMC1j96ElwhGxwFx
2DFzVC4jkfdPSyq8uaQ3ubENTV6bhg9AugZjdIQq83JLpCGaKy12x4Rb/0Zrz2r6++fmwPTXdzRDmDeeOzu3mR23NughNiCT7LmT
D+4IO6SA3lDk1kgHU0ooCUu2HkuHNZIVSXY88TnIMZmktUch7JazThfuFVZ0KUPwZjClRnEoKbPKz0OYbmp/KvgaVohhw6paoNUb
Kjuol9PyMx4PTgtel6/EOmYRacSZt5RRuWROoDTsKgbx8A7qppncxPc8dkb/SENeKowR9YE5Pk3C297sgzI2nmO4evKHzb/hY01t
DWwCMZfpEFWZTNWXtV7rqxzdetXS3B2fwx8sOhEXpuYKifI3h8zwJrQE7eax9RPhUsOsdRokkBg15pfpPKdAC9m9XrJGQp8TMI49
yCTndygtC0UgUWEZ/4eYB3Zcu48GgYt+E/MbKooT9SmbvA8eGq07fTjAiqwacxb/Akvi/k1CoDawy8WCA3tdz5W5VZX304XX5CBx
jXOS44f3q0riOOE1Ju3jZmL4uT7eLxMtV5h8fLmwg9VOPD/+lWGXDuZ0IPeTE3RymQs9QRTQP+C+L/8E+3tZW6F7AMC/uJFYE+cu
EhGCCnL89fuK+LMaH8n05MnqCkq2mPlXJZMFI/QrKHVSeOWx2I3kUmu7LTjt1M1k4IMdBbVjTNvVekS5xwDugSGnjl5H7IWHlAKw
Y/DK4C1sIXs7HoGymVmsV+o8hpFvYQ/VWaAknW6YhGa57kNYFO5Vtifm7qSJsoPWp1ON8FGQL8xzZty2FKEH20Z6y/K8CphyoJED
4nptDuvFMQX0lEKg5ceKK9tueeQ1o/jTGcHk4x/yD6CV9hAlSyWs57p1lBfbd8b0rUFDCopfcRXjnNlA+M55LtGwLaeD3oo6Fl7O
H0uQHTyKgBv88W/6Hn6wLcA+QPNCvHEbIvR71R7Hdu4sC7ZL0/H+rZUFJ+G/N1dsMC2NUs2Z/qzO5mjbwOeZPU3uVTLErOv4zWu0
uz9wTwaVESDTZroXbsJRqNd1SfSaon0tMTV3T1hSTO4BTdg580dhdKA28nUJtI+tZZJ/F8nfNMYR/eCkgIQ87AAwmGdAAOMWcPjt
4CjiurD9bJk+SXDoEAeG5LEKaoE4AH5FWapedpTLoxZXlaAIJSdiWmyGNIYiFJQjl6ArqGHy4AcwBr9t907XTUT1sSNZJRv/jYFU
YpNN0Xixw5wDSFemSSwK1IGcCpKkfU1lju8X3CUxJEYOlvhrRvszci1xITgRtFn2TIHHeCIr9NGI7DwZ7Giz95kAaYGAni3dks8N
A+FvGHgsuZtAjRudOqEB7qDM7b7y/AO1U2gVgo19erOqwMAwZUyK/yMjcBQ+j6i1Y2U0nhMUejl+kTuSa0PtAiBYBVeK/4BvtMJ8
XvBWVZX0GnpcmcE9C6cWpNmy7xM1/WUFGq4g+0X/wBVfK0pB6/XGwMFk+9KNMJoelMLhDbIfzh9gfuQqYA7oSK4eXdSIEnSS7z+J
695MHDAb+n5y+t/3zFAHdhAkopsZCJi3JlqlOUxyXbS+MPRKLT2CC7mHznkjtWkUmNRzwXTIvLpfk9m67evK+8CAV45z813qjeoR
Zf2Cx1nG+1AUqsCDiw1C+hjIMGbpmSRpCfH1VJ+MGUeiM/f/HEJM9dA7U0aIRqgQyMkBUFNctnaYcvpI/pMtamoSHMMRf7hULY6w
JFTcOP1YeEYxAmKnjdQgphzPhI40HHX7BYNEZg5Mft7a8ZyMksKveHs0Mnll9CZJxDsY+j9YeTskVHNyIsZZEzcTiNzrqEUTH73G
gbmzmuNJkZ7tbcArvkeR/0Xg8Ras/A95MAayZNsxqT7tauZJK0fSTx6YTmJ6zCgZSbpcz4GP3hWpouVAU7OVywzAp9LIytSp7ryy
YmyC8i2jPIRs1Xg/QXpwLjlJSxURFshc06EGlvnlTgkHUSHwsYdpyQ1ya91m8jO/XGEzz8fTzL/caJxvK0JTcbOTPQ/FO1+QoFF3
Ewx1JNbV6GVu4cziewgr/olz1GRpPaZIrM9OAMXAiJ34A4+zB9YPfFgiNRcVVlb0l2zr54PcsrO443MKHKKzXT1NTorO0S56xa2H
4HFgWgau4vJgvZbFrMkOH0YOnEErX0qD6IwZTd1Mx1prA+cH0ujZuzNypym79ZapH39w6cZ4+B0MlCmTHhbOB6K2/qy+O/kIUHkK
YeWAP4vjkg+sKuw/n9TN3ULrtVqroH4Qqc734eOZNdf08kqCJEzWYijTKJxRlQKWNl7oLMFkmZfyoytDz9ouND6ulARLWpZG+27S
Vo/r6klaTABp3aECWhD9cfaNguxLJJPG72rwqRGNDugA0cPyWxPG6q6lKsarjo2npptOEvf1LLJuGyHs/xajNi6Qm9/dEIzZbiay
n9HfdeR1X2sp/i0NDcWbKGQ49htyZMVy2gT8OSRwe0W0647373S9QKYeE7EUobzqaGS3aCEhMBlE+bh1H8sJff0U9Ih0dMv/Aqqu
NNbyVQsNtdDdf48HARcz3+uJtsH59/hu7p/PCdwQ7htWDS0zKaE+FzlcOrIGuN9sNMw1Bz2KIxC8z3MVzUVVCHenTpc3co5U90R0
fM45b1cslC2RDzF1jPm9zCRXx4Bnt1JG2THi3/MItGevWOXfIuyp+MMpRDUiDlZ556mL5s/JjncNE+xqUaQIqIzvWFvvyZDVVRrz
kdMQXqIv0oLTuIjBnoHJdur9GP0PDrGZBhbyAmhpOIPY2OXux4UW+8IZ+jYZXY6YZkBiyujqgHwjhE+wbXbBkc7NFy17mfnaxArB
0abF2KbgS2xrePM0C5kqqxNjrCx15nXF7hhgI+NBCQy72akfq2+NAkgKW92Dnt2cIyWttjL5Dv04KH0JbUZLA44kPfbfBrIh2EQJ
GodnpLnCw0yAO6mnfcAteUXL7g6kv1KlE+LrakfhRtpw+oOiPpMHlvBO4YZeaRjIQEGFWzCoemQ/dV6LkW0u/tdV0lI7YjGMz7qh
7vrRpkKWio6iciBz9cA5MScR2pykifpr/OO/ipdpbRCScmuHd2Vx8TedqIS9rqZBSGu6EpFVjVubZL9BMG/w7UOUZEuBgazVmANX
9viQjUy2DhSHdhlVen958MH/kn5v8NP1CFdWexzVhrGOLIAZTPuzQZQ2pwL+nm0hYEkiOCcurr7XR9KUMUGG8wvsGEztvguFHT5C
LjAcEF9x3zAc7RNHo+z6e0HT+njyitkAoYX0OlMVWPQK+NI7mXgPTeyGXxi7A4S0E6QWJwcd0AtsdozcNvrl8u9W6mAisrEfmxU2
7junMlzNJv6FK9PJoKfv7x3BQF2/WFwtZQkuPipXUopykEXSTcE17WkT/RJttYxv0mw8Zja13zbcbbEfXK78LInKYJHdjRze8m25
V3bDFSPtW2i1VRlyBg7Bxte9S2f6Dc9PDl5fCfJuhl/549jDy//Lgkcvl/0BXDXCK5hMJfI/mslFTXxaJhtd4X+rYjT/c4ecCAoX
9CCemlFvwtaSa1XaNksBxSbxneoznft4W7ldx4H/0QQi/ycFU+ePHLGaP78U9un/eMTPLPnkDi11dEX+VQLLib85M3Cg5CTPxC8X
10F8TOWvOakHbORdD+GDTHgOnylTq4Y3vvlbAwMAOIe+q8bxWFNjS4OHPzs67noCO64OEn5Or33X0+HSe63gRP4J8TOXHQCwWDc9
qijTPzTOBwOZ2rl+SfT2RQqZIO1xMK7IXonT6sgkZBX2IWhsrWfLixQF4zI/d24S1PPSOV5QEU4McuedcX/M7rkIM98jj4bwh5Z/
TWvyNAsT585QcgOJ7L1JIyjv+Mw0gzS5qpr7BG+cmnmxsZAAZyVjKrv0/WPsdYpPFnL3ht80nyN/WI/3rx/uM+Hc3BOEfSS10ve/
da4JFwAo8JsWh5w/ZQKyNLZKxnfTkatjlxkYZSO6QDvT736j5A7/Q0SuTc45k7309xhy54jZaFAoSO12wPa9FcP7Wg0UscCnbP9i
RUzFWE0dDuMTMRwNN+ldq8+kh1/pr44NCupp/8GEuwL2Dsuv0seH4fh615DczmdFEdvJDartEKkAo47A3OyePFKQmpemxEb/Ukrw
TXtnU/+moeqDcRlerbRd0kWOmCZUV/3VtnXcLiUBQZTjDathSBraHLTwFdT1HqyQutl7ARUvSMvGJXdIfa9gTmTzalbtflycfU6k
Im8kOvWJpLixmMx0jwdTjNqUtcDtD1hZCw5ZpAmg1BLjSLwfHffPmcI/XyIuHG8/zj1HISD5gcaRiVOvRVRzYQT5uvccVfsof0cn
hvRCWj38/MwUd3lv1cvJAMZFsfLNsLTvemtHvdxfjUqOLt/vVw06bHpzTsCb+QlGzAC+VRvoYRkPsL+lmAcQTcp3RHNxoo0NQgP4
R2XIsxeJCuhGGRTL0GcgpkyYMJ8SJpQatCZj3683wjXmrFuMxgHSy4Ive6vSpB0/1KR5wCtoPHRd+XmN9jQAgqyEczYKVnCC4f0U
DBOm87PYR7BoMhF66/lnnPiWREZzyN+dNi9KzelCznnWJX+ryHG3Cim898cw2ch27vKLRX6Czife9Ljy834kLW/lr3sqpix3yr3q
FJLosrnqOhTD9OfsY5ariSq+rCGI8CuJyodd92HUti3+08EsKYJxbGzkdV9RBEt6xMUTavMxYTSqFVHmibzFHvHRKOkpzpOJf4wg
Xzrb/ObV3uzKb8h98rx1czbeT4Ldv2Z6ZJiKq4X7vDtd+eiInr9wQrlCsvl1stIIYqycQ+9DB3i7b0clxiEEekbHl2K8yumtEvh5
I0uGZp7PlUdTk9gQI1ae5YmLaZrvds8J5EOfnUB4O/PSB4HsZgXo11xf1Rz+o1mAFo02rXebD3IUJJvqPrD1Uo+aEslnKr+My90i
G0a4fQ/jELdvr3JXQMXUuStclF8oFYc/d4/ffM6gs1xZIsUiAgUi9zMig5+6ewy8yFXO4vvB7UFAgJaVNaYsyNBHq5wMupJyGVQS
bHykIRJRvVOV6rgbf2B308hA6q8+/ge4eHLZ2rnsKt5kkN5kWi0N/ItEqsguXYvoMdP9xy5YJk9zwIQaB9czNvdOcskZIJ90zJd6
JYD/eK4o1OdWoQTw6CvD2feYtNXAQzSD0V62sakLCMCesB7VdvTKUP/wZKISt6qCR0FXoVBA75TPCadkoJJK4zh1GC2wZDY9d0Od
Wy4dFEWsz3NckxRHSqod7SJ7Zn0qvbRlomkMq/K0wRNS2QC9L2HJXyjUb2Rh2ogeSuzcxsGHld9lLTUzCYf3cE+0iSbLe6GXGQ/F
O5oDrlud6S8wOGZlqDynhUgCJy998vCrw32AnuPddJu3gzf1u1MZtxh4N/o+yWanBuXZWSlcZKek4hU23YD6klumHP9mySP0TPRZ
2omqj/lFr55gr6pAJ/agwItBW6e5UNWoY0VrMscGz9kvu4LqTQptIIdGXKY3SPSSLTKX541RfFyQ36EGi+d6bQ5us+wLApwpgqSN
s8WbhcLkd+LI/2axYenBS9pJjv5KmzkO8Ct7LxdcIzb0gBZAYX/ZCXLlIiOdT9Gn/gSm4Mthoj55jXGkJxKWXrTIsUbm/XiWfZHO
TnltRiQ1F4b5t5vC/cF22TbNPa8dqGFthkNpzH0DTl2d+oPRo5WwKTeUqNMHXMWzA8g7QiFu7QkVke2ElHhu8flaYIQTs3lpnRpP
nZqTTTq53MDz1VaWuckwfEphyfCbYNIUiQf1xMuBd+IWCJGpEDQzWv9tDS2ouAgvnwThz5wmMRoFBhYV7BCrs/7Ff0j0t2fls3LR
7d0dTnRBEkWzj/y62oPpYXNUuQWckIRV+2Pp/idEip6zwph3CgN5xzzTR3sgJxpHI4Y/UOyCTpsSgkiRflhGajwlZzK4Y1yBM3fH
ss0xJmT+RsyusHh41fTDo07a8tnKzygQAQHpTdHXBC7ZI1UdCYu/Iif5NHKUREiIaWvrOep30wpPUTtOSc+2DG6rxnWpYlhehDIu
OsJfyy5H/PJiy8rODtXfLDsfhXtBuU5otW7JNDuoPPuGOFKMy+mfoDllW4TfeGFS0+eDMhrMUgMuScky3jCvk5j/62Q27o78Jdx6
3N1SnwBwS+TBcbzLGpZMAccX0+qthGihxQEJ+SUn4SqQ/dgE+T3QfjJ4iDpHejISnhO1/7THjYUGgKpTFK/gjuUE2ITIKdEIe34C
dipB+eDyWuwjlaX0aUWkjcGVHMAVBIVh3lIcYGbTt8G0oByug/V665NZqzKn6dRjLguHw0heDmnUPS2A6kq4TQMzwnjFNJmIE7CW
fXxTkwKiJF7hKnhdjINXlP4xx53H8WMFuGWuXEwH42affWPGNiwnlT1UUwWrOE9gzuuNEcxfKLnqJcK2SUap2xgi6AoKQOfR5qcL
bkE4MP6/f4GI+3pvkm+MrE1yT1IGJ3DN/OGmWid0L0NifM6YZPYLXrbCPuG3rjlYEJDkMEyIvF0frKcO/dXmXuF8iA/p35+Sl/Ss
eYOpWSvcmpoIZ3jMQfjFA5p29W98zqdfyEtLGhCZE1u5rlE+8kz13c/od9e0Wt16dAYCab/WuJE+HUSFl3R9EXvoXu+Oe6jtcMwp
IR8E6bt5k8B6jwUA2SStIBii/awl5bE2tl1p+a5n7kXBA7waN2D0v2cLc0M2nvuSca5BA39j+Ew312gCy0GKd1MfwrpooI4RbRBQ
Z3UD/VXkzCXkzw2HeY7L6IhsWvm2Vi3cJhaXAEpkhWjjsT4MfoBT68+jsguoS3zZKjQ3/BQ45G5t1rmVHMlm+qMzMvlDMLeQ7lCQ
GT8iimGhVx9Ejrz/qghqJs2GgUpr7XgxpHyVpKH8qG+GCNX62PBXvDS1IucOvWx5QjClUMlv4w9j2pIDIu1X4Lxqy25LH7qCiqix
EwbhWRxmEjGvqV/BU5XALb+aAO0RwQ2jWgciC0ya9b2sGMzt/23mRHk30XTGpiNaBSYz9Y6+mZDFkk5M+WzfQxXo6QkfLfaKxVJL
rNNDrhINTNrb8CsAmbLnFdrDg8lFyQTXjVj77DBb1PNeUgDOto2k8fSWuhNSKPp8iLtwBEruD78ZQ2ALfEGVls4W5ojYW59P7cQ+
mUbsL+cTiF2hAisMrwN6FhLbuTc4VfbxG1m4cOZ8WIG6WNn1MbyQBGtHQh/zwKEeS4s4/9AxK7ZftMM0pRgMyN4IaSltzrW9g9Cx
1ss6eShFzEXgF/o6NeHmP07wdNBqVKpxnUATkV1kcA7QgTSi328M8nOkpcs2HhcJIOYGIpmjh2qp8/feUVO7j98fPBWveSSL0te+
8ni2H2dQDKYguaBcN3THhuDc2dbzplxXEbibV7bJUo7Msf6Jr0qkKnRmY0jQ+GPSvwNA+jeG94MpFUfxmMO7q/W2vrmDQ4LbzNEx
9J25pd+hW92OWGfDgLVEJadkjk8qWQ1exilEOgLUHZ+4dEV3/P/qdoJUO4V0D3X8OO4aa80cArvDyyfDza2GCF6XJIv4fy8Eu0nW
9L14K7U52zoYLR5155ODakmtK8YRiHJ4ZxJEbQEeRAelPB0abwK/a4ctfN8y8f1Hg75ad0bI20NJCITIASSGnOJSSQddnJTwlMOG
6kDk50+09+hIhrihwj1xNai2lIUBJqtwUqsPSumyDyfzFpLcRULUfe9PU4o0/UCIupXhxB35WkYYCsyY2oGSdkR+y0p8y4MBQHEk
KPkTr/lYk8PeTo5aIFTrgjfGRlz0nZ7y7OzCD4daZwlDqxD+PAzNLXWhGbCY0SCh3DVucA9wfoa2TcJO3oq0zNkXyrdxjGwT3PVc
V8z/jDxj2eKjUEa+ERsUxEw63kRmeKt0NihDhoH9ZoOGfvunXAxysRND/WqxjWtgZw73M7NFMbNaicdB/Ozcv4HSsii0Q0awDdOt
nrforsglnVwuLmsIHMAyuSEHjAsAJBBDmE4BqVMmtqN/A1kkAGwjAe32Rq26a/b1KpicqLBN2xG77z+xtClP5TEVOT7EW3wZAcfM
IugSlY8Ze8LpHcGGC42OLpD+v0XDszMiqK09n4jzVXLAmcuwJ+e94nDeVCJZDGMsooR1/VSxNDPVPmZ7HSZ942IDiV2Uuh0JWknR
XOBRwfaUcIeChEsHrelFvI7WKDy1z8L0MGVPkDKzazb/EmpQiWDVRcyKnaJZdfWvFzs77Amztzz5fOuJwypYyvflC0Zt30hpQBn7
4/l90SsfCIARwJVcytfbr03dhmHioMmcXOVWKFW6XMhJ/szuiUv9NfuTPWZJiPFSjVOVHKATZxQeqYbZwOBxT5Vd/1wim0sIHp8F
1qehVPGqpi5OEZ6EFMMoGj9M7v9d3IUYkN4pCmqLF4pO7rhlMAQvRw8JaOBdf5J4sNHcTnszOdMfcwFZKb11/IhnoedM+eUEVfUN
CJNeu3Zy7ECiSAYtbcxASbxUi7dMzPsFfpFei9Q5HhzQ3lGfb2MMckROLKQdkebBvX7bAqdLPMvsgwOEIrP8FStMNaYKxzu/J74E
VFmzRBmVVWqoOhIgr9KEnfxCdL7Jb9ZhpjW0/KrM75fmT5ocRbmsqaq4bIcBC+Spq6kyIvlMkVzuTxvMOM46EM7qOY7FtGbEPesx
LKMcfC2Fkaqy959kppDAAm4ADz4KiCkzkcXYTT0cAiw3nVG1+10c+haRkG5iGL1n2m5cu73xYMO+lZnPILx+IZJd6enHYQyI2s63
TIexYz3G8gJRInr8ZV4badOtq8gPDfMECxSSw8AenDKx76IZDsKwWrLN7KMC03BUonhT1YQYvWjtLujn2ji42AU/BAYoZxWGtRDI
mD/C/rhywhUb5cU10aXBK1LkXhq5dI3awv2SLuhmSdtNAsJzVg6NjOF11K2i4KPKflenK3JVRymKmSpZ9XLeqZwG6KOM6NK9w3LJ
WYzxyPFfv/NKPGW7hSCZxPaCTbP/OjH6B0Hxxu+Z988Gc3nTrppbb5cZ/nKoJQ2tCDSjn7Sj6qmXeWh1E3AYtqbMv4jkJRPkQVx0
DJijJRVVU/0ywfitNRYw6lIj2JyylqCTYnRKzHOc4262n5+4SsgoFV1yOv48BcDbyNxl5HV5y+K/s/qqPD07p+WLj7B441PkaeTY
AztqGZp5NdzykY2NpPFxs7PIweZ0oWHFtY2XbcVPTrM1vQL2xxSNPOcdcrAdvaB/Yrz6wnK8Hzm5jD4V8KR6rXxNv6dwHdfYjkgj
0clS2Ny6MSH5zcL75pDEYA3U+GwRxGcDikuObtV47/lfpJjJGpSrm/tU7x9066Tc4//vKrg0rq+i0lFPtkgZByOlP6vsxqviC4UO
QJILVnKiZ3EJ06tc2NNm9K49jf3SsuY9xJzWQ0t2BA+4il8JGJCSkYJzyROhapXe9qSOP1Wnuzs+Ah9fKsYHTFUMSKRrzvZHBgh5
gR6HauZ9i5vNGeHki17XQibFOY/ZDoD9r1Nlw8PffxtmpTjQM6nVWRZ99rEdp2JdViy4SAdT1nGWiUjek/XyIUT2yFPxXwyVNH0r
8Pw6JZk2uSyvhRbBjj0Z9Vmp9XxNLoq7G+i3+qhcLWttMhwll3+efO/zMKljuVpUFZvqj1h2I/t7WFzLrumsWAjjZ8dGx4wCQHGY
VTcGu0H+0ilO+9qNLrXveWT7jZf2nOOdg9CSagKt82dOojYK1KkGEDTNHZ4r0rqyMY/UE2UgonEz1FwgoafkXtVe8G4Io/rvRR98
ufKYE00aSGVRPCa6AMJWu+vWEusVuHeKs18ebsa14Zzhh/3QqkUbGCPFelQ2BRjvWntVxSIAa9ZBStLBKmhvJcLSjy1SVnPLzacL
BU9gH267gh//5iUXYTWW8SgFoyO4se2T5/A6HZrLf/dd7aC8ii0xGzu+ZP4omO84pu3czwvVw49o/Lju4RMglJDZE4RX9MxOfl9D
nPDd9F+Q4vzM2YkmIUhjJK1vlYJ9iNskyHsL/JY7LSDqE++OfWggonCFPhG+nwKGFgRwgKvQU5qHgE6FqO9SQZCPZQuusjts2rO/
ThXGvxrbXgc5/sqjBpaxf69+LYoiPU+Mliu4GkgR7ygIXYLHVcVaTABlrHJcUc/zyTrjwM4/KAU5Rfv1vUbroS+5RE4BFsTBM8JI
95L0LHsAQyB8RmfW6hFs9Cm4XMvxb2eeNOMnXuqD+F4lKU7riENKWNA65yasLVgTBOpMFe3tAvxNIKVEuraxsDoURNzHZcK/SlfO
ez6HxbHhmvjLzvPCyoOl8eOI9Cp3jOT8wLw3lgxlZFEJ0nAZ+wZvb4n632Kzv+wP7ZmaRnqDTIwVNkK0VTmeQZIh/IPPrx+x61Cy
+x12Qtnh8hPKtoeXp4FbkmClZGl/emlnWqCpmLZb8b0pqsmj6v3WOaCFZnBIkfARadJ6rF/utm5jv9VcafBFLEuvYQxKLlbFJHTx
/Wl88vR0btxWESarFDbZimRBsgRxtVxMTnlCzW+jmbCKXZqH4szuQ/lU/RAvroLzgVkvR8s2Euto9QUROwWjnKacVaK96yZ50Dv8
A4aJU4cOwTr5BviEo2dwXlZjOHKz37/mR4CsaKfj25c3zn/fKg7FkdfsVeZV56ZFfT39tVlp8uh37k4VfHVf80d9Qx3jPko7iLAu
DsZP3HzpuM7O8IGZQBp+ssIgUrqQpGeTn5NjSW1s1flW4h47B0SI/EDsx7hpWGzfGJ1XhXGdkr3L2g6Rli9MBH+P2cdS8QQPZk2D
V8CKPqX9LJGQAIjBtvUKb86FVIwwAfYt3Nd2V95ZlxWEuWygq99e1VrUXxDTcWdhR+l5EQvlWZr7vbXM2XMl5bLfGKsvIGshyOPh
QPRV+/1y6f/RW3cWhB3t0ACzP8F83C+lnqgQgW488Qsf3iS0pzVRe+rzSFj9JSGEe0DD0au0gVTBASDLRVb9axokd8+DgHOX3gna
Ww/GUdBgpvrosmp3f4mgjBBkMpNcT1kjY+7vISdYbP0TWBX8RRnNOurD3ddDhwox08A2Cp/iqhuvAzq48T3PyWnSrFsigQZTq1iC
Pvw4m8H/bG3aPZ54lwwXr8kWtJON/B9M68jSbguesamRm64EhnM0zprjFkTShh4iORvwt8iIimpaodOjpfN6MCmuORWvcMHRAx1H
EkC1T5Ei2FYI0lUfWmzx7jayjZxU8NrfHQA8m0A+9rBO40kRY2VJjecAt38Xex9LctB/09UBRMHJ8LI07hzA86FFbFdl1foQew/g
WKXmE2AaGXBqQ5Y4Ja8j9N10EIm3lbZ2tWGUYAhz6rhEG+V/Xec1M71+n5bj+9NkStmVqnuwiyRWzjE5pM8gh9Agcx+0SelSmXtQ
/y3JXDQjkNwFRbbeG6s+mDRGv/iLSoP+4gzQUDoAqgt7+8pqnCz8DYnO3GHYpxmaTg8vaLojGq5Wb5cyq9fp1bpq/4OAaJDL985M
eTcbSPpyHA6PrRCZhhgsnhSKa4YzobHViHLrNOtkqF6b2s1vkYdqSnybARLRqXIVDJcg0C6kNtg6sbIoOfd8wUwbhwnWBWIcsiV3
e6OFJTB1PkexO4IoDleDfumYAyV3yFbnJpJYnmYs1XXCkspwVYH7505AtguMmuvR+sRXuvmQVPU5FdoGxGfiUKA39u1LQx8CSNTM
QCUiBtng6ZqzJPnC6VW1S4cBB5cWIOe7JbJkCkWVEjnZRWxurtrMWRVWp386RrXHJEL5UWgqSJY6iUKbYpI2ufIsbmRvmvNYbC03
JgU49LM0wyGSenUKgnIjheRCtyasaZbeUuSQQVNXKPlkyOMxjHvZlEu2+hBfMG1WdkdRxHT0GZV1IT8NWW2Z4heKKHZbN7ZO6caP
QL7c2aktuwWZT5E7TH/m7nQpuU9A5iIOfIeVjkPFqbi2cpbZrAm5KDYs9hJhaOKqeIehrQR+B+uxWmirJmtwKmRLQxD52jKhqdAX
p9+MphUH4jEjys6Fu6r1rpzZbv2SCEldCA7GJ4eKmrtMZTVHw+TkHOk/ZhFL1bKaV0IUkggUbnMAtVHTXGiA8XP6BA8/6brjUD7W
NSOa+IBtXi/CmS8x0sK/XgcBxrq9UczToPs85KLHhLOPSq97SBwEY9bTtGCMsBArc2FqdIzg3Qv4CK04E8x2zsmsRpjBYSiDNlSx
wQcKuxK7wcTMO2PdRO6w0RLYubU5Mjfzsyb3g0j6cgg0pulQspxzIEuq71rebdv+I0KduOCT+QYTsGVrAbX0+pL2HV8aDesby3sq
yMjjBkCDwItAXZYAqCBUFyxVRXnx4G5dv2wryxgny5t+XgaPaxKeb//sLspams+CbJNMEKo73KP004nKqWtDK4ku77FqxkbE5u7g
YNfU8XckwyM12vsRcqwcnw22K2BOpul4TgE0iJasdor8MQwgYqnVj2CtKaCm7+B619kKA46GR5eufkdk9Bs2MOy43R5JqrXM4Ydp
kTPh/CmH9QheulEeP6pKN7+EuwCAc0siGwFGZz1T6WwVXZE3nJ/u7iMPl5FIbUsSDdILAKFV/ZrTDJLocDWpwl6KVKfU2C1Mj3Kw
oHvKLJ6o1IDgu6zoDiiDMhab6IkiaKUS1uVd6uVyqMoaz+LCHvnrbLorb9sb5Y1R0GBnDVhFZ7LEw/9zKwrgvPj0dwCheFQvshBU
FOe04j5lB84wjmHVn4AAzIc+0LuUBYuvTUyhQPHTuR9zJUxLpGJuT6rp8jwRNWyVDtqFNZxD3SDIa1NmNXmtngw4BaLLgVXKcBOt
JpIJ5ql5kht/6abrZogWSNItk80B8lV2+xheSFCePd+9FAqHEIPz0+GxfWCP1eDNnNnuasIW85AR6zwOsnCawLKKfTi/brW+5PSn
ouwCA1XZQQgQIrUmEtfqAZkaqcm/p4ASrMAuoW7QORNJyf9potyhHJzY8KNOpdHvrTDsovAisVm0U/QYxbbNC3APHroEX7wBgTrU
VRw1R3/GigMyjMc21JAvoBs2lwmKO2lwh0kC1LhqZoW9vqPT4nDAcq8YNobxdgA+g9s6aZmB8rlZTTAZfim21DIj57WuabVk+ctS
r1bhIu8MUnQ860h77jeOkjO9rbvoMKSPuuY3x41Ax5DMazyjm1UpxkJfSvLzIPl9d359610wJJknnvEKML7VqWnz7/6Sa9EXDpab
ENtH9yxYzAGbNP0GFRipPj+YP454OsQk3TBbPhH1o3q0v9Bny9OzbGWoa6+lmMyapbkcz2DtDT404/TK4ic26GYSsNJIWocFVQwd
d9Nc0HOUyOiM6xx3oWjiT16UuZRNMbRuUvpVY5bLwLCGt27T0AX09v37SZCYKGWn4Ym16klSE06i0uT7V0V2q08VIS2+FB1kBqFt
5iQmCNnEEdVuJMz4YXTv/3pEoK5Di41eXwWWpQZtpsLZ87/sIaN7STy7I6JFE/HffLrfP8GZ1XAVxng1UYLdVNqCbd+KmrlKRjlj
QyEdhNaMdwBggcw1I3jDGm8X8ioCJbKhIO2PRRg3XtUr+7Zgi4hYDSDVu7GcinvADERIaDmVEcjbf1035ZulTdUqZSZ4ZCBGgluE
NvsMBGWDWo7ZfQMsiGHtSW5VlKNd2SsCqzHPu/xmUZiGRCuwhROTatkqgYdqdb88xKrpShWRBv+em4jRums2o+cLFFhTx968tcNQ
BFY9Oe0G0ORz+yNbUBxAfAfIipfASN/RxJXUdna0EXJVO9WCMi8xSikcAe+LK2eIHtah6EpDEg5ZdZdQF0rxruR+m36lI8/8s2X5
d1RTetIM7h0Dn5EwY5qx9sd9UqJyn8LtlqvWqs5Ofl+LNrsx2WWHcb6M/X1H8rraQgZgMNB5C0FNQZ28M60OQWj5QqiktGDuGz7I
etkJf7rFi7j/eMQRR/tbNBaP7KEbkR2vKOSsqe+IwlvRjayIay7IyaCQOy8P6RWHY1a1+yU8RF/TTUTo9lzntG3AZuSNm93d09aB
K/5cwF7UUvEtG2Q/26EPtJDWxINV9Kog9NVCw3EBMa9Ene5r+Z4EWEMfqrx/t+hFjfPuS1yVHan3ElUv6mdO20233NeePuJm7UZ/
sQe8rOOKn/re79Nq0EZfDNBZu44LEcgMHYguyqzsDXof40Vy+TYKeS/etdOH9EcrHFo0aFzv+kMW1qQnTWpX57OrLK56d7DvJeZG
hH32BxQdD3KXfKHQSbifu5fMgvgAoFebq2t9sS76XYwz/jLSEcvGvEzdhzm/T0+UtROfrEALEAD2dpJ4nKRSxD0NRNhDDOHzyzws
DwOVIL5/AYsssrrdfELA2pfs4oqwxobrVr68QU8/IRurnrnAsrYxlOcoqjz1L3RPe85BNKED/PgmWdYu6rAVi0Gfh5FdHRt779hz
bMeYftPIFB+9jrXEIeybojoCz9jEIca7Ytn9uavcVXDap22hSLUXhqjJRWhsgMVUOAPx/NHxjqi8pm+5IYKC2NsYg50fODXyIYff
BlNSNDVetQj8Wat5B/m3z6Gk1zwUvoA5lMUmXEumY2JGqj95yIJqS42gx9gcBnOqH23LTWJd4u6YD7apeyR6cLxIxW0gwH++jyob
YA92OXqhxttDEog61MFYwPVp6kgBPp3aVyGCHKrqIBN9AJB7Ch9KQ0jQjXeo6rNkNVJgqER9EsrNpNJA9+z+XXz1u/vayMBndRpN
VHce4v5vz2beRp2SiyqsqRNsnOOQ+R8n/jAc5D5sBrTQ4rE9RFQE/kzrhnGS9aJ9WmGmt1a0Em2pBQCwZds5DMhS4jcA0oRQyegI
lnEaFnTLogWvwGtUU57LZrGyfZuylz3UvIA2Kb8e7v3PzkjePB39rBEr876a6BnHxESDvc/T+RYQNot6PFdkCoiEC2iUAu8Ic41/
q4SV9H2xrVbeqhHmp/Ub1dD9I50AeEQrVNgg9yWd2KpC+ctXr8J4SNIcltAZ3r7ThOeMNVbjLeklVb2XrzwiZ12qKU/NOomKJgAa
diQLkUlREULANIPOnbXe/fIxonsnDiUO2nOyuN7alzweFKXU1eFnjwfs2VlposYZTUZIVkTCLHSzwZ+zxU3tFq61GOpC8p+pOltl
Dh5HEpO+b7veC/G3VwXcTYazJ+sptWzTUqYapNZUrcDEIH1Q0CG8gsr7y14cuNdmls2UN+BVNqw8dSMHtImpcy0IBxKXfXqo72UA
k/HVwCFIKZFMyaS6un+SYiDaXTsWzPE8oyAVYKLNbsIn9CS9ewfIKXUrPV+p2lfb7M3lGNWsVcxJYso3+ZS96SaHQnU4qgM+Nj+F
LbdbpAYEn+hJuF94YjsoETh07OobpkInDZBu/FopHpSpbTDWXNX2EDZFfGVzkzh5eo+2ITMdtAatssp0Ph0uAN9rEHoXAwaP0nFM
l9h8xkHoF2PjBZ7g/uqa/PISWovsPFwQN/1KmKERLD+NQnfiEtQ61lT1J4unt9+CnxBJ+uIALE9yauiUpHrtbrEgZdETIi+MHEAg
BKAyR9XEL9lF7PmOUeaapvBcQzoopz202FLkZ1Xfy/vtXK3Q7XWv/wydWDE3HJhzSxvCTE41Xpt4Ssox0fh+PVe8ouBdhQCSUQPW
Mbg/7Q0Bi3OqEjotrEijw3uUkE7EHpHCpVPMDUJNPgtRPNYkR4PvKBdh4z+dnNTggTkgTc84eArSOABxt8JM+2OY5iz9xZQrzf9X
Eit1Eo6VySSN/6BjWeCZ2dIs9YhSrMm/NIC9gMKBzJ8g3bJ8cJ9WtCM2MWPX0xMFqdohEl6y4m2+5qbBG2i+Vt+q9XUOXWSawQtY
h+hY/pK/+/4iWGn2SQcG/hRJfGyp+SjU4+1PRe3Md1QsS5hzq98ckJ5IVgZdY3e+XG5U53eXC9A3GRiPRD0UCnMEv1hpk1yM13qt
/XB2i9B2vvaXECQLfYy8H3LlIc8g+l4U/tyA+gYgTK7l1TGExM6RKpbRMgsLGbDiGdVwKN4R9kxPeGVzYFNcMuQis/FEt4GazuqF
18FH2l+dez4gpJ5mqUJVFt5pwwieXLS6IE8SDQ3ZTycZ03EJlywloDYEZd8BKN+CQzDtqQ2sL//2eJkiOE96sGA77pPFX/1Vskki
ZRR17Z/48uIucXtILu4W04dfqNTriZ5nq7NnRmkWcviHpKEXeNL/o0lq92Pks2SC7Hh9V6YM8HNl+1rbc/hAp6rKvnmO8zBYqy/6
d/DI2eHtzSk7SFmfj1dUdkMqmRZYsOQXihJUxG2nq8ptYRRiBjlizec3T6UXCA+8YtE3RCDdXkdt/KunLbU1K8IBGkQNUEWH+umg
yMvYKOeZy7gDXvi/8+nJzuSbgoYHn1Ozlzqo3y+lS7dq+MR3B2cgojco66kjVbnilHVVZAu8TCByOko32sWiBXGb5wpYTL8iGGBB
xfG9JDWnENItJu6O6gHEALc2hOD7d9FyJ0oo9o8U5SChCOwDhkW3ibgE3U7IV13iE+1uk6i1zSmJXlMvzPypVbvGhSJp5E36WNMa
gXXgRYMW9HLQZ7kUtaqdqluxnC9+auSZ/7MglrRNW60SoiuAkGyYJyyAr/NjI0shbucO29og1AB1IAwjZ3CfaPcPBaP5orhxRcBh
RRt6iWzy+ayMsGVVSHME5ysp1+zYYTnVGldpNcc6owgckhl5QKJlWlaSkYL+WqVUF2epmBuoqgVG6d9LydE7b9juEbJn3zxvMz7D
r4ARjmR/hVpMKj3l5UKyackmzH4CXF4PKQ4euVLD+OyT3YLDdgwD9ZRA0fghBfeM61gMqqU1HtecwCbzhgdz2UkfddDFmTaYO5ii
yYUFIZSd1kBET8Adxh2Lz+G5NGWLyUPQzBEmz+MNbJIAGox4TcGd/5c0ckhFSqAaxttK0nI+etpywLB4+vTqtcBT+eiAzE+lrpEY
YYfficoK9RklAdU/4QKrAVOF5c6/dWAmc+L3FIo6q4oxpdt/7YDJnay2QGF+Iv0O2jETrjbFs0LfeGiH0aRrdNA4NtuLaoU8cCk7
wdQqjugnt0tNOzgCovQ/0g/whQTSiTqVrd/1MMObSOi7n1SYBywIrnJU7lj2+B9HUKpi9Ys1Go0+lsiQu6Wo9sSjruxTbxND9470
0u/0DY0PqSsHaUfcicav/kgpBi73hAJ/bpZmKLFqtxBcpy1ZNqXptv+ciaw5le6I2d5yssGD0008hVofHj6o8uMmkX8l8vP8QR4U
5U+QcVVUMdRKpvA2Dhg9clHj91994P3aGqJja82UJhXG2eBlr791GQGaPE777RQJVl18bpPQ+KFntyQEwyxaYSrBdqm/H/cUoLwj
hz0ZgeQ7DcgWnrPx97AG57JNh0iGwRmjtBuxb5q75MYr0sCdBdAiMdVDZhR3DWZo7x177GQykgWQcmaR6afwXzmU5mW1lsNamhs6
IfqYJAJAxe7cTBHPMAJ+nxGXYborAng+DwbwoNZDD3a7uV6uCTexi9RGMhoW3m0lcDax1zDvkwo5ynvlIXLW/ZJC9NanJcmpXgT2
86wjehlRl6tGyKJEIQLTwochjgOMfVMIBh5rQQ1bcZHuy2en3D2n5p3190bZi3YLibRy676HeDvg/54mOwAoukybvPot7+YR8diA
ukCS79DLdgQmj0QSge5JHVrRMx6EkyQJBV5Ph7SQXfSqRlRc2hsGPoZZEofYC8170Sp6WZUEp21wA0Fcv2dstngBEtDnuSBcOpRM
asCByc0edXgpDYPaBVx0MRAYmsEFIvkEKcJoMc5L8pWq1ueBrq3MM0CI1bdeT+47lO32Eo9Z2fftajMnG05Lj4UcEmDcZ4k88Ygq
sm46DG08GKBtNcEYCGRi4JUdReL13oi+W1/4v/XNNTq0K24/3i4BwLgzqrX/OJN5wrzyY0scU5lz/BrduXAGDeWYAIHA1n6M3qcm
jcHjbb5S5GCYyWu6Z/JDJFHc2GllBKI1hbh0Ziqzk5aVhanMLMw1UUbyIFN0ww2LehNr0Dhp2fgLn9JFJ+cFN9RNrBgAq3bhIDMA
cvJh+WVXLASqYr9GFOfUbdHjCxZXfDg1XZ97rMEfW5ihvf1ck5KE/A9PwjxG5aGiz4fNrRMWvAa20Ljhwb+CSDDILUVB7EYM+0uE
wGcgDxJQ43xbD7dDMlfxUHcSj0o6vhNDAhWIh5uU9+ZlUm6JVbMwAo3wxc8GeI83Y7VRjw5dc1cpwDFGSrow2lxZ8f39OvbrVZY2
Lo8HccpW7nOwfc0FzjsEeNfr1GWArfobfWF2ZRmHL1R3pCnU4CbqUXTJU70ftkVjdVhGvuv9VZB3XBMY6F4acA1OVa0cs4xHvKJY
FIoDGZ55TrbXCebJYrOy8asTaXc17wjqRv4BYQHYr+sNtHYCZb/iOuadsTscggpgwKlby/5yD1pOrPLHU55kYHmogodgCOQLa8jn
7S8cflRW1o9075Dq10FQmId+9nm9/vNg2iy2HA4x2XzPp/EBuLOqpYCFbX2t5NLaZgM1drexenOIKJym9X7WiXG5LW7iUGqFE3LJ
DZ2WsAOICwW0eKS3zGsTEpeMMudBli/JfDapHPN8sfdmZ01yqofC6dndjZz2IuNBcHphxV1te3BlGh/BEoEOKZje06RPNMHYP/AI
SlyxW/VaLP25i46NnLBIQFBFNpy5asyJEGg/OAQVr5UDhAEzwQRuEYPqqgG99XcsComutOntNchL9ROGlIXJ6qgORQBfJ/k3RcBX
GjB+vvsgTMzweVeaCG3cXpN7eYGz/QB/IvT89cC2FnspA9ZJQ3/JMGTddNV2Fs1JgY72YmbaAGaDFSXRjPHi43Uh5a4fa1vJZnnO
fji32u8vk60/0h25cFOy62yv0HyVnJBTUdTYC5Lj+PzW8z60s/bcZ7N9U+6+Jl0+aUXgo34P8RHpVdPyl+uG4WmEomuNu8dOXvXa
C36nTpWQYWWD9mJmqpYeY80SUL6Cs1+u7xnhqH2aZo0HpSwg6IIqtjN5F0AAxkDbNiLiI1l+fs8N7IBq9SSD8BzA6DXaC23CnptS
FK6G9bEt8GG35rVIhMUGoS+10HEzqYBrZKTy4XYf1Y2/UVWZSeLJH0lGrNLeo/h8G+eknAdHNu3+OcuYUZGy7GityejEavOsxA/p
HZpnXvWeSXFo/c5g/ZiB97Q6e5tQDJOqb3GEdRaeGMiTb2rVW2PECvCd1/Dm7wkqLmLxbjxsdG4+kjf5jqK7DxH1SrdXffEv01MP
srm6SNwfrGhZrGgPL7RfOW62qcHk0Q4YTRM67bCR0OTrom8lSlkO+sRlqCCOP6Ne+4YXFoZgTX+YZsAKecj2681msUWnmW3mfYQ2
z5lmKt2j8vNh/MJOvx2pLTb1VojSYi0/Oh2BZAWUxpjQ1k9YW255V+Chm6g8DOHIj8SHfGIQhmN7Bzh8pbgBrdSC9IwDXYkFZKRw
0WJKHfs7QcCXe/t+f0fr4d1k5ch5l46bpMi5EYQtFuTfYyReKW/QtrtOHoxEtX1B2D1Fvj48BkwZNDbCpjQirEyQVMa6eq4L63K5
90xgrGM57PJ7vbXY+K7oUpleUp5LfDzVFyi2ViEhTknUU0YdBPKlW23ItT4SOX5pFcseb/Bk9qWApfEvkhaFGTV3ilnp7CRhCC6l
Tq/cdePevtC49IUt6b8q6ZjkK6XPyMdGGbisUqy6zZW/39wkBXuMiduhbRX7EaKKzChcTiPI3E+AsFYaA7EE7/z5BP+YGD/ucbhS
DWermj9cUC9i1fW28LCyAEd6FZ7q894cFKqA1AmzwV0Xj5K5sOco3HHHYYkEjnbfd4Cdzz/f6YcKRVtwsTEWAKHcmscpMzXrHLIp
Zcp3z2G/nftUNUwzvag5p2zClIF/Ut9UytehTNxUJwQi+ef0QiMQIjcLzncWRbcN1dmWq7Y4YDpDbWqFUhwvb10Hp/34m0WKa5CY
g3YDnqH9Zi1YDDdrWnW6awxfaGP/QtJrrgPNxe/4vqmzCk+1s8U3wccfTjJiQmyfgNpBy1RzoHTLkH8cYx6WgELau+Y8BBZgsnVg
UhP/2HuGuoePMSqVTIorKdT5sdLPPEN58pukK7oDHDtZls8+PthgO9QpibHQ6ANIZGEebezDXkv13TSXMvE0AKUMGfhuY30aqzNN
BqwtEY+eKsHUslFOQLeJ6fUYBBBGkEQ0HvyCHyX21r9coNZET0IdsNfHre3Z48u1BiU8WKKxv4ep7b7XuqRXa+HPrxljlrMXaCQU
Et+VWa2XXe8t43mw4xAp7ahuc8cJSDG4xVpEvefdaVtCi9utG77ldJqTIsb33Pla2QjiD8HBH24GjNEzplvUMhXw03rVgS18LRx+
toV4xg8kP9/8KJoC6H26WP4PTUjxLS9eb6NhHVEVNpINkCzsail1z4BJeoWUbSdqlHkOK0B5CSmVQo4xTVY9bCrFBuOkSeg9tipG
jf/P7wNrh3plM90yN3ZlOO2bGm/9iev5UE4iqIcPqZON0gC+54LNGZhrosbbx7tPnG1qJwTV+ms9btGcXayKWbErCuPBp8gsIXsA
L3URIVRY/0HDO03RVt9i8dk/vidKcO0F4JQF2eXdtZj6OUGi1mgjLsONsdqm+oS5DIm3oVkfBFbnxjUey8wfOVKk7/cVVyKS00eo
Ne8QY8n8Gtcpul+EjK2nWdkDPFa0YxfPXYhIEux/qDF77PurUNA4d4lj599z1BlOFdsPgpVGYLk5B5WFhropyl5Mwq5tx0q5Wf8Z
MC1v3Oy1CtxV35Er33GBnHEbVALknmGOEQfxlpfeqSTo6clRmvumOSkJ89jFcOsv3pyNV4Z3VkWFuHpkqOrwHlGYqTBbSyYq4V4j
nCpyptY6vtuwpZf/N0SF8495TaNj6urw/h30PQRfvCuyJKSdkJo+O7qjncIxYIJNEIGnRQmEiNTodQegEgY5j4CaawOzhN9izEtY
fiMcAf+blcxVkMGMSFnMgqBv3ix1S94MyIRqJOvGUsnJTfTAzf3QprwLvKgWXpEltYXiFZxGjDmn5zvfEqxiNrjhaM5NVI5lW5Eu
eAksZLez1NxwVE6K7rEQxjomfIJmmnk3OvaQ1Rhz9KdRan0vflQk+rLCytjSMNOQy3TawZJ4OzPi55d11DB9ch0p2Wx0jQ7HGLMx
lwrOnoakI4e/PIRAlY9oNLRJpAL6CJC1KLHm7RXXSH3UoB+C6nPX12hBBrxL3CIiJ5cE+UMgGNIwWWjhbdJmS1/c5UVDVQqAMgUR
rOl0ouOMDIA9wNEIHhPx8M0cPmBtUaXCt1Ijsj/xmz8g8l2ltXDd3Fp8T4TaMw9YwiqbSRKE1UXCH3dpf4J6RodvyUqUo3gBO+GE
fSv6BV2Zp2CLFe2REb0H2k3NOHi3oiUh+LXb5povhcLdWCcrf11+u6y3ao1QYAKChxtOGCMDa1yuJWDrMjPpUEeJrj600ppP5FZB
Tep5P11wqP3UVqTh/CvVSP1yHcfBaWbeNeyCIWSpk0ytIviQYCikZ5v+R4nYCrTyowk/ERG6kSU6BqvkKYEM66ipA1HnY5n+aL6F
ihYm3TjsAbBNt2S0ShYNwuXGpO/ztimZTgaXPILjL+xEg1hMqb0+pIcq50TaTOcn3ULwYBHYxyhGRLx7v2JvLGmx6ZsW3I3FTi8R
LN0kH2o95f1EaOhHSkKKlvbB3kKAQ2cG6KHQL0+FeVrT1gfp7S170spoHLaVHZuWzVprulsH4AB1UEKUioCWSY43mpQeWaDV9g0A
GD8PcUVrmj9QKje3CRZf2iqLP/YyLJcNKiknhQ0kb5rIYzn08TfSfLlx8WEsPrfpRSMbf2ygT4DBcWiCoj3FzeB/4wO1lyyrRbx7
AnKxQ+dkPdmrHYdcRpMGUAYHI7Q6Jl4yyOdARK0iKAGj1yb1sxLlaFoXB8sp2DclYH3GOFM+duDqDOszpFifCXRLwG58CVkQ7n38
6gdREDjk/UPpf+asG5oMDBtC2Ee/ctd3y4QTcWZLEDbyjTwE5+ugQZYIIjzkiR/t2p35YNUubMaEvTmkIYziyNIpM1NsolmcMTLj
r0/gZ3g9Gsz6Mke5BrZ9wCIs8Hmre6ysdYDfFPu3ZftwRqPo2DiY+F9PEnFJbApw6MwK+HZKm8vlvPSpWTZ8QvYflL/kAMlrejuL
6m/Jf5CLLS5c9NiD02vWygZ73Y1h444VRuh0UtN7DL/6Zkf/2gFB59Ojras6ThiGLX76PHOAmGwcQj0Iqf9hznh4pEV2V/adsUJ6
0YY4wcf/XB6/j0pEkehEQBAWlZ0p7rnB51CwbOjBhBXL1hZxyqDI0vOf9RrLcd9x12cEj/2JRvfOI0EncdfH5Ao6YddEl1f+OusO
4DMNmNng298ctY+8mJnNoMbdQTxaK+fycauz+gt/0w8FlQDGA4+An9Y2XcFpbe2gTD+W6Ckw+H//qhKt7tasHQjHEvanLLcrKboo
Cc8yLG7B7M/yoDfV+XbCbBi+ZZ+kz3CFBBuRPzScaz3Hgguy5xwrFd0xSR/NqgK/oijCJ783U4p/kVuQBctUl51FOjAkvyOSXWup
T+hxwRcLKb9qwFYaVpMRfbcO0crpjooybs5i+/hnr5ulkuv8lHukyBf5XUuk+9ODJwWqMVDdPMfV9oueoOFxzdGmgvV1wKnp2+9B
FAKbyOwcrbCB0FkewEYPIthr6lIzXFO+rR2nnHazcG2PMeNSohWp2zL9HSEE5UL8dtKNn4m39FB/b9zMgqq79dOTMVXXOBexuFeW
TcxTQKUknYneLTaWbQtBKIO13Dh6EFl4n2g/s3QSY/I2u/eNeM7cffeHC9a05Sq3Z12EXiqirn7g/nfNTcJ9MhI2HZa0nRu811VT
GUjE+RD69O1mnSAeORgL9J3nBTQD4bxkTcRfliH2WHWfNGijsdWmtYVumf/OsJTQ8BVRuAegZwhPrH+Bwjg6F0ArddqUWMWgEdDg
zs5HeZoDw6lUEuK7WxBcTeN2VMUO0g1t5Gf4O3OeUOfrgAn2omRgGv6Swsl9jL2RY5kyRd44LAIxGVNC/ogDcGWq31LzhR/xt8zO
aFx8AaP86kdXYzjlIVVBV4T3Z4ATOU5eg6sKVepab4r5HS+P/zzFoo7BYum3YMO3SUn0ojYGPsrvtkL2wXu0i0m8KLYUbGkg95K5
GDdTiC9UFrE1oo9g9S49PUx/K3fPqs17dYAQdvgsJ7pGKx2viYPdlzzWjdcJiPqV6150bg3P3Wv0niuT2PWRJjKCR8VWsI98pQyh
zcT8qoGPutugs7H/uNEFjFy8DFg2vP/GD87Q/iFl+iDeUNFA3aPrUpWQXC3KtaJdlae+F+dXri6VGTnzmkaQeLRU0b72ZTIj7qdF
ARIrzvhUMDc1RKRuBa74Js2kLe+vKbH/YhoXj2wCpOdDHCMyU43f4FhDVNcFGBxAizQxGYfV1nq9Z3eM9RTVKJyF0uro3NIPdCvm
zNvV1jkK7KxcT/nZJdmPYHM1PaYl9MLyMhwOFIjyciwQpdeQQrqCSFXDCJkYTrQWKt/h30axwW76djuA/QpjH0zTQUWokg4sU6d+
GVdU+FlSwY3diQ3MFePFwVMVl+V9DQkIrV5bWXjjFNpGdKraI3RMAdqzAnKkpOc/ldej5nn559LUFotb5CgO9IbQJL7iKf2VwYr6
ndg3CPl/yFBlsVsM2vL5/9Ssh0HpFIyGPMHrHok44Vg01/tGsEji4u77XYXnBlDWBc9B/wcmXrjlgvrECIMu6JFdUiPZV95zpv8v
bFx80/53mgMNgkkbQo5sNCFmDszGo10iRbjFrAJ3cBgqtYYr44AJz+W+6/DvaOVq7yeDRcvDIlyD1CTmlUc/3xU/6/kimUdo1+Dc
ZB5OqUSpz9KTQh8jl+atg3aEaeOOaGQVzqtmmpZRiFp4B1GASPmMc0jq9ZrShsgMbNWY8UBgRrFw3jlqLVM0FCpD8gkbQSr+UqWV
/kg/KN6xqRWVFoR+dhBZuYCuh82N9iKSXmSH9vgAgLZQblhU/6Bea9sGw7S2HhBHo2PFSEjfylOPY9MYAQNqcSELCYIvt2TQnaoy
fMJ5J9GBFAZoqiNSxEOdpbWniWKKpB1n8PV2ApmV1uoouXj8Xnm5/PCV1X4zsB3QJE/NnZUoRURF6BUkcO5AB86itu7c/VlYj6eJ
edWpnXqSMGO9gj05mekuo9hkLFDHIb3uXaAxZq0STuVE6EBN7RZLN2kmiFVWG813o+fWMQM7YCZvD5uKbFcRT9PhQdhxzrwpdRGX
e5Glq4MbGbByKILW3b06ThMc818OQ/WDQGUTo6EkOnn+4Bqpgf2GLLL8HcGwt04MIDu90Nb6VpC6wZspqIORUAY0aGukA8ra36PX
hF7jNZaNeeoO6Jne3LjstMMPY1Ie2RwmO9ftnW8NdBlIgczRS4tgAFXrKbLFJEHVo/G5rb5/zGqyaeI6j32y83T7jT2RJRCClvbG
95U4GTUMspIfkfNJpzlSrZrQDDPq+IOP/WxnLyHgwm6qL4xPGFZ50EYzPrPGa/J3Js1kyeQDdzQiURAQrQgRF/NusiKSqrK/pMvR
lP4ZhzGTwhHGmQqD0zARInq5TjqhnMdT/qZkT4gb+VbB5ztSTSNqLqfvG/UOm0vEhttKdrBxcBOaSSU7OzK9I6XuWdWRes2B2aXF
lGQHNX0yX0NqxaAOn+Y0xVcfJnQBJpVfrp767NcJ4KLxW94Nmt2KhMZ28JrsMQv22CZdqdWPwLuUum1FrkeLL6d0ldu5ssWX7DCN
yTDpKzFeQAZk1Q228aJNTEmFPRwHkR0NzfUQP9YmEYFoVyQ4OJUuGf7fKLsju5+UdcaTxAPuqtf+rhkKAhQg+7IHwgzXW3WHaz++
UhKG932cEinWLMZ5ejDZa4RWPNjDo9wh4blAwrvLo7OvZygj5HNcu6zfEK42Yd/QqGE1TP7nL0Q75opP1its4TIKbzRa4PRN0Gyy
/YIfwFxnRWGxHuy9UDMAj9KmmReSthRIS32bWpgvx/+DIXqYaiD8RNX095oJyWJIHZs1zcc8jhbMQNFFMOau0ccxH2pJ3RiJmoBX
zuB4a1c8xCrXwAy569KM2RwnsX1OGRCImzXoPdGKN9watvPMEmbsVMcVVvQ+9kUqwynj1UxLLwD97HMcNzLn9lHeKyx8lSyP3+f2
YLTNPbNfLMHqkP0/OUqYTsmjMmz1BYG1iP1esljM6Qm/eEWUx8mCMhUmlCDu8ctlkc59ycdHt5CJqHnyjfCvxttOO2SyCbW5gMFE
IBWLBTLmPks7qAIIVxsMqg6FD5dnGgTn+R3nRlzVSra1ss5yB3iG/rJQ0hz5teklIW3xf/DJW6pz/9yUT3bGnWIE6zj/YU2cOUhF
//VDPdjmedZokmV0gR6Z1ogrT3ooW3CBNtFhwAUyIr5XAL9NvqoGnHgevhDcyVV9L3ynkXqZQ1bsoZ6wx82zjKyJQRxenOyWwpjU
qec9Sl8hlHqhomAf+eRwgRHl3neOU1YkACZVMM2tFEcdyvj1cwxVBk4zqS2kOhMQVHa4U3kQFvzogioYMedO8yfjehZYpjMaGqkM
p5cnqCBn9BuiXs8lKDSnIYTILsnXpGZFlcjnR32uc6BAVZH5FqeOwZ7vKGUnMrDMy6BGlTrqYLgc33yCAGL3chGgEESRpc3qg/yq
eKSgg6b6l//lguilF9ifhxPrCCWL3AeJ2cZb6zobMLnAxt4ZMWfhjyFtwfbpuViInxUri+X3RoU5RRt7PLKBId38XjH/9usa55V8
MW0YZ387sngNOJAiqqGTFh/LiRKeyeipwoOrUe/iokJgJ98BPJCbB7qRDJa2KshKijaqPoRspOl6E8DleOyL2v3eNZup6YlV50KH
pAuxSVVWhOvCsiEy2La6kexcmVnvRotYX4b/KfYq7SGi8nre5PoGU3StIStJald72o8JHogDqTyQFy8XrLfYPuWUt6c8q10NknAd
oQbiQ05sZr8SaKATUJZ59A40f+dA0E0wlpVwNIftb8ICDTCW9oogE6c3oQI19zGaAW8/y8f9sShiWDG/v1bS+vZfRWB7lvxL76Gt
foyUycXXZ9MLhG9cCRBOMmVZGWqDxMVWCbnc3Xe8t7IXIlDNoZfqP42Hkz9kl69JNOHQSGzhzA9/YINNyd0/25trsRp1UwQ0K6/X
Pew6L0UirBjlSYD7r7uTKot056ntvUV338TkejQCAXsHoy3MTVM/w0vCBKJnuqkEBnkik3XAkaPEc3gHuQWnH8m05IEFO7JDn8le
CHXYe/7NGPRoOlYdJG/o1Y5hyf7DuwORbhcVJznUn2yqLbxgrp8x61pdEcBjFhMWdbPEJDgV2VRQ82P8gp5lsp0EPiJeoDMgExcp
wcoTHuuylgYqUR4DxGqRIl37+9w1xDN1h/qXUL/KHttu7uVS0+8OjzvVPnxL5Qg/YeFETzez1W25ygRPS4db5fSRR3DZ5ny+a2z1
eHm/lpn/Qo2F+Y9REKV44oqpG39vI/e+o27IWWNj8Do+Oy497xUMnq+dXBEPjiKla9v7BpqhpJmdJv7b0r+BSIR5mdyJc3JDxyyw
Sb/kx5E0DKFHGybTE8tAnu16+BtZCnYX81HDVaeKz1pysaxqdXQktVKIZ62QfjiUgnnBCDwiQilIt/EkaDx4gShlVmZv3wKbmRSr
TwOcM5Didfj6sLvMhfgDZlWpwjnyQAGY8uy/JkHXCgBdWFBnJSkiyhmZEwdXXY9BxZ/RS4IItrd7jtIrPQwEKNZfx2LmiQ5O4lDE
L6QsA7w92aJrOv8XZiSjj00/DrT2Js0YiZuXZpXF/UJURFCPuEQaiqwgzbdLcjs3ry+6JT+zBVvOBIaCcU95NGEyWbe4zj2Pc4pL
RXwFVsU88W2C+/DkHWiQ7JyHNSTfVvINW+dTfdiZdHmlMPQZ+rcC5l6m/WJBpV9Z9xu9vR5CQW5j0Gi056nLIgBb/x3vylHkLTyl
jpwNarruYZjB2eDL+mhjIRnloC4gClblgfYbNyXSfX8ihn/uASrhwSkci4hzlWCqPrDGf3nGISv1Tbf5QZ4u5MucUa0z8xKwGGIY
8yPPJJrNmW5q4so/iifHOwFtAXmrdXS8IwApMuFMDp8Hua4/Lf93IcvUKq76lZYBG1rCD0m4EfyaLocJgbK26QXM9Ui1XUXgmnrX
ZKGSh3FwnoTrqOi8tPxUf7BPPf8TJfnf/LelcllNzuMIHcr6/U2mMEl5EvsGDTj5Ih4BgYC75oSrjG7KMbTY+CiVYlForjIzub/l
GDvaPyxLM25mVq5HS/FpM3DMcYmGjof3miExr/oct7vmJd7vTE2LYKufj4ESJH/yzAXCKZJi79JTiboO2px7p3iZ1JLQTuR0OMFo
cKidiszGUf3/7OpViNyphwjhvGVu7JfWg2oDFvfuyRrTjxg7UdnuvbZjv6G4CGyE7s2qit/uOKqsrnYrQ3HSRRObS0pKk9xcHW40
nqGOl3mx3sxbXOz0Rk7LC0/UaeXupCih5x/yNbmPCyFNI6x1YBFxzUvErLAF0wVaCCgerFobJu8sl4NDh0FNnSFgbLbaYM9y7VB7
MpvexvDCAe4JfazG3AL3LYEbU/q9l+zaN+4YobSwPtjodB3E+79Hy3sQkBCaLXoaOCFPYDwt7TVP6hPN0kx2JtEwVrYskE1HYuZr
FZgW2O+FZ57cK3L3oyVIc7rafZD08f4YTkXe6JEi4Zsh3rMV2rLX+PttP/2X5EWX0u+t8zVvCTb+mv9bZf+44ZQifI3JGIrQRvzY
zsU6S0Vr24No8cv5ASw2A0iSWzizEI/gihN6lZE5S8hexMZr9YpKAhQs6o2AjgpKeBM5JuJHR6tYLhld9uFwAt9dlQIlH7S2jqMz
eLKVgVkin7Pg/eJ2YfLiP1Zdzs1jcn/cAQ5EwXpd01iUhysY6qxJs8Q06UpGmxGBuEOx4GC8/Qz8b0M5iwghv8/lkpJ4zltFGWTe
goi31MzoCea7kA8lSnHLQlO6skvXqYfnOU+b2rNW2J7HAZbWLhKAxAA1J60QJGtHnJs4+OV9QUOHDdUhKOEzldeODDPo7McFgg9m
9Mb9L9+1WQ2a83BDI/yRo+LFOZ0hGkh3ZczDBvJ2LlOdTzFKchaty60mugkSHByPZC5aWjyw+6ThQhTv9GmgEGPi5VRXGHIX26NG
xUzTrr0mtfHmqpFjIjBFd6MRNx3Nh3t03GjBumB8zNclFfu/kGKgUR7xfZi54Jp3MbUXLfpbHH5nvTB35La2BsF2hHGO1nKyUpSl
4de11NgXGEl6djDFNaFqfvpANKJ3AnHc9hoX5mIsR1+GxcTc0SDcRX7DyWdo8w5B0oVd6Y299HpA1y1jUtJahvz8KaF2SYgNlYxZ
jOHV+8gcwTzB/3yjaiFh/Pcrbx6erFQ/U+rZPgt17+e0eHbpvdSslrmeC3qZF9flWowoF4NvHxC8KlRbxfpHpDYV9WW+GyQYrNqf
eK2qKOFUr1RFnw6MenF5KmqNg55bXEkeKLrMqkUWx2ajvn/ASdmHwvP5hu7URueJlUgCNrNpKH4dCxo/T3323Ibcj/HXjUvE+KsC
32yLanmMVzFY0XD9U0Dx3o5xkdotH5Ed28IF7WT8pb/Jv7X1sdv63o9IZ70GbRBrn0NpNreBf+/TI1TqqNzAsz1meFbr/x5um+dp
3GcoEkJrXCDoUofSYDVV/kQ522uhzLTezT5PBlLTk9BL75bi/JYmPhE0B4wQ0xW0Os1RWiaFCQzwqk4yi/pK/6OnAk6+gRsi6dnQ
cDgs2AoB364d+nCJ8Usm9WKhrY2dirIPfCfG8Q4N87wtGwlmmIGBsQMb5JLEpBJHfpyxEA8LkfUbRAOaQKyUU5UGnEo2hvlFccrZ
q7c9XX6NJApf+RdL5ImpQuprziIHUSlQ+NTu8a4yfYf6PHmQJxGPCJ9T1j40yEbg7OqAkgHNl3a58NpbRWcYzWItxXHJRENzpAid
kHr2U4XobI3LJEXdUAbB1PmevwAEiPdGeP9RoTaRVywT3DSB211XeOurl0b+76ycf6Pxajv+4GBe2EY3+ctijricT5O9Ul//MTRZ
/SaBKui6iv1B1D1VDy8ejyVOq63bBhPVFjmUQzsqBwpqTlLpRtQr0s49VDVqgBdHeGT+dUWdWhTNwfsgzV7G1Cs+uQEP2fZ3Tlzp
27OVBcpBXXxQlsLIMisLShCqfCLbGLlF8FysUsj5Pc32BdeEDOIcr0ax0Q5pQXYJy/XqQZUgLLgzHn0VgNGrmETfoLHMUI71/01O
pFsEDdw6l8tHcHoWtN9VEP65shxIcKqHpTsRGcSXqHKpDGbKIionxQBITRFxp3t8afmK3QVX//BbfHUtIggOREm7eN72o6wJGh4X
CXA1xCZfx9szD5g3iHQm9xwcwOpUrlxlYgwfYTQ0dIARU9XiolOku7MJ0aQCqrpIdwP4i4BRJvbG0VW6nTyrp+EQVMV44BGPrqbq
cJBACU1Gvjxm5Jph9q+Jp1cS6XXEat7Z8SUUJm/RTYWrsbJYkt/rSucDsIiRp3H6t9QDzBVJW2kuqCf3ElHigSmhzUrx33bjDb4v
Opo5OAjNhB0iOasT2uAQbOMXpSSgvE71nAHASu7UFkzK8LUEG6L4nX2QU8Nb06Y5gdKjZXQUG4rYwnGfJNFsY3KmNY0bUFoJBwSq
o4oNBRopvemv7Xbzli0Na0z2ReV5ivvKzZj65dG5xCQYTlyUWimobVgDz/8Nv4cfLsHn4evk4XBxjDNTh0rQ/fOe+pptv27hbRuU
6cjzdisSGon/w4ZdYbUQsBHolp0C2rxg3uNrDVYqg2U9bpUCRBQ+SKXWEgx6/ejIaPv0lvPrrddq528Jas4qR4Qst6QnMg/638+e
5m48zEi5mW62ruS3NaWrD2B4hBOblCieY2mVc2Qb7XgYfHPBvR6a7YCX2hdZydMIfKUP5z+I0yyJn3iReotziHv9L0LXbvuXkFpQ
BJ8I3fM3e6TAXlVzKPBEBrggC7ExzVJ1aU9AVe18uvGp9Ab7PbAwJl9oFkKpMN5hQM2JYL4DdNO/+X6+4/JA5SHGXfsyjWF/eJlu
trOeHD4GmQ9X1lg9UpWjf7rB+UdGMTwxuDUhnlq03f+ihRKjFLxMMUi1wygJq/eupHHMuXxf3ZNoaDdiWqP+JKDr7IlcG61J1X5T
L3x6Svs4DVImqTLQbfGl8cBh78ElK4E7UqLSAyDcjBvhCzn6ky/QxPRYtaTO2l0Tdr4IUI5VgFp+iGl9bGjVuqXLYhdLA1dZzpKe
DXyxYSq7d7jLbPFjFysnitXvjrCxKJ1PkA0XDNEwjKk/Y9dcZTRuA9qrx2qUBiinvIonq2c7FMQa/Xb5o6g1B5MyBU49lAr04XNJ
2bfUe60rIgL6pWCrnDMGHf6Zrf1gVf6/U5SoOfSLJMv2lX/wVSyaOexlMnYUOZOniTrZkoXpJeCtXmCOoxWTokJuKJFKfui1A02R
9HcikzRcTF0801JSXHbI+FQ+mbiOdrG3oSiyK+dGz3uVOYIeFuE54iPTX59jAFhBoqJlITPg6LMsytwzGLS1T4Ng6v9UBP3AgDSn
NXWqZnGUAN9iH6+zggLfsUyk7oeQPOzu1rLEwQZjckzN3YQnAKIgY4pR7oeqRtjws+eNlXbScpGeibCVGkT9oIIL2QVyXLV1S0fC
drG1Nm4Kbx13IMEv9/bnq+9NlIGsfSLB/QcAP3arnu9TLr1FOyDAlEfoKzjBs2T5r21T9CLWvPofF+Pux5OlpFgxERM1FMy8jLmn
+HeHnhV1WcX0RSUkwlRAb9UiHEmdmeGd3hpcxKnnIwfVRR4HBiMBQwTv+BgsgeiMhRpIbpmBA61kBMYBEjLPD98tF49DdKwvMxhy
/YovmvPPfYiUwz365i07JLd9WDd/4S0RphmLp/cvD0JzPllBL0/dkuQbGC7lz2gLppH/mFpk5nNDKTaeK0HFhwLWqV7hJ/DuV5Z7
GMYQAOYIW2l8d05RrGTSxhA5rLI43AZkSREpMl+ddKZiKcjXViN34ZEs377LJ1feqeNEOd3DqNcu4Z5DBeaveHfJQOds/frtG2dH
WMfiOYUMzrvCfbU44Gk+6Y9cdT2MP2I5NkNxUux5gyNnh0aqcmzUbEyxz688neaciVs3oVKQmnA0m1aUBe4hmjuR9ASGiKN6Zm1o
znYHh9giZxwQjA3WWIVwTgSqfEHWL2RbPP1pJkJGTKe8X6vcdeUM/sifp2l+iwTVcVW2SRvrdH7kb6z4eelUe1E2CAxkYY5YuKKh
XWG3v6iccN0q4pYj7+mkhsAKOkROpPkDJtEon8gatmG+pj0MPDJAVyqYLhERhE6o18lKx4jA38aVq205O7waBRwG4ysMuWKfdfVT
3CuRvf04cwIxp/s9AEUMXHSbMXfQz9ecPmHYd3iwCZ8X01MFiAG41uVIzUjQprm6kSH0evDSUTmpr4FAs0ZWehnlia6bqLmr4N0E
ZqUc5uNzj8G1PX9V0EhNeTuIYiW8T8Muojg5fvUagntjK6RZYHqZ3u94RhHeXVjnNLo3fIcH+zuECtI2dE3wGEW1+lxQyN0V0rIV
u8m1Kfm9vheGKHIYtYS8soW3L2wewVvvO3bLOnSHBk88IG8PMFcNYrdsvCcCFAb7aEvAc/HyBvDGCXR5WVcwfAH2SLw/pD6lAvXZ
tNwxuZDd3LyTJkZl4kRhHGp97mqsgmc4c92nXwS1cX1hYYe49aBdmHHzNf+hvg8XinwNafgkx9C4A94btzPY3QjEGKpTs2e+Mgfw
iKq/tRnisUPt1jm7m6JwLv2CQqGCb+ecjYujFRNJnbqo4h/Hc7YdJrN5cKC/V6NX1BL+n5oxQJs1mzdLmdyfBxtr2+KAicnD3LWk
8tICfBnrq38j/pINYPl8kb9kaAFRIgh09CC8YCybaPamjg3jezItPWNW717W5AZXQo+xTJ8sE/csd+S3ve0a9+IjtAsfQ8ZZ0xcr
4egP2J+KaoDqjeKZGVZdIuBS9qZkoS8J/16UjPGGE0MCF9Q6n2tGH4Rf6PGdSpGe8yACO58BYpecjWAY9omp5sDKqvlNgdIVtXAd
XZh5YMr+qGsD4p6vdtc0AQpA39asPdcLilhvEWH+ehNhWhevD+bCx01xLkxVAs6RUHkh87MxHOconPlnVKJQUXNjOKQBaaJDEgC+
oPPCeW8GXgW95s4M71HVuppvu1yO/fmayG3/sAgR4EZgpDj4fyq37NZYjr50mzCrmYBsdvCFxHArzOVcR8D83w9tTuke+4mIBl5i
ZOJ7hgHyF/GQ6Xn9Mach1lu7NFCK4B7O4oilfz2akXRH6WMulIEinuolHzOjNFi/MipeV5hPlT3i3P0ygF5n3H0ML+AxZzReNCbX
YRQX21Duevi/EZMndITbzcaCD2iJDhLw5YFEaEOs7ZcOyvwm2C95/zS85JR03ICv8oVNcK/2prva2fDiD3ejur6vpLW4UbU5LjJ1
G6AadO8qtrbSypZeSUuUyM/3j9/SUQcVlglhTNXtw78fNhWr8RHnKbkT7GxSDzHOYuyDAex0MYRmZFJZpGh9U9H7g5Vn/j9J8HMT
7wir88OUITz5lHQtJpDuIn3RwM2uOHdL/ki9Z68mFYevAw28en59q5FhmDnZthgfTYF849yU9Hg2nvT8sv6L6P3jKIIZCaNsqZu1
D99U0oi55EuKh5gnsFG921dL+PJ1Hq8bbAT3+UdgMB+BimwuR5zV32w7kk90MMAaf3si7NXwjH0vgHSJZ79caiKws3bTZutsuFeX
TlJgUOygmbHpg+pKe1qyI+mD/4RlSlStIOP4eWyjYjjqsUFK6hnrbVvcgDcWOrjxHyqPjMNjDQxe1nVQ35UzlDtU+b70ex/XKg4+
Ekndv9fsVey2Xr5k74ea3Z5iitWt7cy90Zv4ygIFd0VRgfL+Vsn4msXOOS4tWfAZXRz1mTQ53teVh4wyvaCtWSAVzq1Ed8Ax47A6
zy2kYDl0NqO/WcZQ/WCy/JGSYAxL5N8JANIFjkFym6j1hDbOV0C4AnWpIsLGol1jwFWD63XQQ+6OltlJhfe2qhiAAtcbkm1/mtOO
as42xrt4yf7BLZYwDeeLzpfUEzytKKtq8Wfi8DQ09PI5ysejWRfiUTvD86PKMaTGhaCg9ykDw9BdZeCePLTVmRQ4HZJnqNT1tbyw
J8673JWLEXU8vZjVeRcUei8CZH49Htpdt6hFYvgaSlne8QTruL9gTDMJDUkvMimQzQfitedr6/rDWDAG3+LoNejgHBmxKe/nFHCQ
dW+aiZYri9eAPufUESY89w7gcflUEJdhB1RDK9pbFoQkCNGbDHtO1rv60Eh2T0BDx8AflpGEvz79lE4y+cPt4CYMj1Ogqlp8g6KC
XkAZ0m5flRhl0hudI/okVnCIW7fobf3mu749aHUoPCoH+FGECoaRUJs1UDfiTKWBcC9TFjABxea1i+CtvOWNYFWC9Sn+ES0UYhK2
M6NpcheEEDrlDj60j3yTX0aPhwbmWdaeCvBtA1zwhLOnyZ21Ifm+uW+Kvmj82LM704v12zxd+OpFsk3lhfyOVKrNBnPGQgvBePqM
ZcNygrGOgHjiOCnW2CX0tGxX7HAaSHhjkvz2tSJdLwY0xNvSZMhCiiE24V6ukM9Ab+4G8coKn76WaYg8oq0kQmsA9NDf+1byQaO7
vDaacJg5r6O9M1HQsDAEFfduNQeCP6N2mlRLKlJPXzvqkF0x1G+Zebg7lCLhtqMwv3gmEOolN73AURypgdUHXjfc5ely9TSiwHM7
6wUds2r7JVThjDbec7yjRXSc8J2lyXRHoUO5ldKFUDkUkkz9Jx553aCM8yj5b3PEuYtS5Ab8FPP2aopJuraMP73LzWOjMbeC3/LQ
1vPiva3Z2b7g5OwRLg+a0rBiA278KryavR7OQx1T++mTD773q8jPYzEuoyD4y6d6nKNP7XzEVX6PR6CC1jHrpWcBtuOY61qImKjM
AZSZ2SS6tECr+T3i90rO6FrIADF0pQhOjoCge12GH1UG3Qxw2XF+ZC+CgZSdRxUkLop7V6j2IYVc5PqywDKxo2s0y1Sm8+vRji87
8mEjt2YyT2uaVYfhWxxuLhA6UCc70+ylQawhwmVinPYwxRhSOrWVdXSoSnpIjbek9A1o4SIY9d5ffV/5tFg4eSBikZM/onxdZODM
/+31YsmCtDrlWMcz4tzi8Q6QAanQjZ3nBgAGJMygAUmLuVOxFHF0eq3ONdSX2V8Q8cmmYqtFhV3u1pnVp2ZpLYRm5Rn7qr/GJj8y
aogJnMUyd3/Tha03189T08E5/u0mH/haewqullw6eR3zifO3toFT4XzCI/JH5CRWcV7YqtJpiiMB6/ydc608PzOYzOrczy+gSYhk
gYMirNvZtgzJiTZDFelziyILrGKX2+abYkNC843dccAGaatLffCoGKhkRelgsKUNWozybioWlAgRpigi6Oh0KeSR247U+N1wgr93
gJ5FvyMXzGdXHqaGuWibyhcYln2TqyjjesgthI+jwG5BAcMUbtYvPqVpopqEvqMWskzXKWJwn+3gsBEet1xEP8+1ocYCuDmh3qHO
Zma5GrPSLINXPtvzptsq5201IYvaZ/hwRrBI2iRzgBtvbuwJuDrpULadFb7+AcB/yo8JTOzcwp8xVkMfmzaPwBifJHLvUi9kvN9I
eLQIOnNLmWzgpdKvrp1DQW9tmZ6OsQruiTdyr209Yiiu+AKV9nrB6WhtegqqjkkbypQ839/6a+6mdzE5dVv30RnTXGNVNJ0y2wsc
9OTySzLWgJU6nM1jBLZ4AyX27PI9yHG7UHfGNt0jNY7wusMNcBPo9toeGaEsCXAVzpj6Nr/USyyM25ioq9ENltt0LXc9CYxA1lbL
1nX6A7bUSU/3lrhGtVSXgGdmejYDr4U8DbKbl+uGPWZH/jUu//H91kJvYVAJ6CW/mjtO6Dv47thWrwuAyqGaZL8F+m00kzr47Ke5
YKrqKbC9ZGf8gUDaexPjNaD7n72az3+EO5fehhV57uDrbVIfW8BnO2OleENRaqFcHazMFbYXuMrvz3adHsOsgpr4mxBsYCUylRiv
/2bfH9ctXtbaDgRVN10DHFyLs7GFN8kuPQ2uMJUhgPx635DYivqggHA39iUVFfVJnYiGAIVk7Gz71i6uwZAFJ7Pe1K432U2/OU+x
LChrlELklQBSd/Qo7DQ5zcLM5h6a8ahjarA/IOyRuV4Nx3bNCo8n+xrZwRldVMv5rTK8Pnw5xhWMntuNtJ+hpb5tuui1mp8xriX1
VVFwXPn279sSHtWN2R+Ngw2VKAUuT1K1WDcQGFMQE1+09uByTjeGMyAqBIOFkp2Wwr0C4R0JxyR3ptn/K0Z5ifjgNMEOx0ylsCNn
be2IQ/Xt0XfBc2X6a9LPv4oOdPpLTjivuto9lE8Xk8D4R0rcKcsVaZ/froBc5uYPcNlZfe6+ijhU5yLmTMaHgWFejS4QpBTbZ0wI
VskUC23+bAGg5wXUXm/2NO90xfH6RkyGgcIKVyt0FAq8qDszVLTgXiXHmOVt6zOudEqL13EhOTBaFLP4GOFGke2QbJwNmqtZKGBW
dTHXEHZFMcf8NDXO4UNCmvJNZQQaW+Hlf6/t0hZRB2omGJaV73LZ1m7JWJOYyT41ALClvd48IBP0jY+HqvjOTHhI073YsuvGA9eC
TGQEOyeMa0ee79INLIsORaxWDPDKh8l+KJ1aMK4hLRaYr5Fwn7gTrn3RJ++jncx4wEtWmMWeh2H/1mojv2h/EC5vIm99YvB3lA8Y
RNOekAM62AtFBloS2a26XIN1wzJcGUECUpFLWC/LlZe/a8urZUYcxSyZMBB90pVpvEiLGpfor/E3DbKwjBkOWeB5DId/poUU2QXY
aThLJnRBOsn7Ku2iuDHZhrhvqEU9kPUpgZHPPWdVOwamX2zK/hHFytQ7+uYAozdaGJBkzkix6vZsCXIwd2pHksRadIVjKswG0TGY
/VuXzdwmV9qAY/g28+zZKRhU2PA7Nagrxx3rUerQ3grGaayaRgwlT5SLPi5F4OJz5Hj1pdafVceRrKpxHDa1D8rthJnFxS2HFyLV
0ATV/cGVMO2pFFaUdGt61Z+YMGXDOdzgUbt0NsJpVFe/JdEZBeq19GdHLQeKVOniCMVux1MXJDxxDplSV94Dpf79+GFv50xrspim
kfuqkSqExV+actkfIn2euKSK3BGjEhOpaG/zhAIVssrlQ+GE3RazZhkaYng8DF6fqA57N02G5DmqtLShSJQfdMVzRHUTtUXjqe0j
NeHp/c77q5LJ/yGUF6XThjvymwL8EeeP5AamJ6eoG+mW0wA2k7Tj8ppbSV7zHMlubc9jm8qFsH1cM78aGx+pUAmRiN59wHdLBU1S
DYyk6SfP29ikH+LNyLy42DyLvAStdd/4m8NsiC8E7zLfqXJXCv7Aze9qoveR9TlqenPi/K3EwdFCLPef/ssiAFZWBeRCnrLoA3X8
mq1gx1/pSAaMwny0ZS0Asd0Slggk3GN5NwBC4tKOrFWKk56luHHEocBHTr6bBZ8SNnECI7YqHZEZUWuZmS/82ygjik6hOkCbRdW8
mL6Z9LDtSvQ+55V2eo7kJp+XoxGHVq+MinX6suwYMeV3oqc/UBrQ3HZ3d4NN/0bMzpW7E8S/x54XxA0lbrDqD1i8GBYCF/YctZHt
8gqxWOwTnJ/Lbka9AKOtr2m2f2TiYjvQiUSoPOjg3NH5YS29AX7C7MnjJ3QRvS0/c09dMWyZgL+015dZ6o7D1AhroNXEGmOF3oS4
2EHj25iEIptxFCqiQQlmpmwCDzCMbAHtea8r0OAcemOREa+68ggOcJyfmnDjvRInnPaoDxu7saOS/3yomGN+uyvbwvCWUWaYNMdL
Rc1E/dd/OcfpgPoG4TDVx6LOHFrf+t9P9/H12mr+pdCEH+0ve4IhESG1liGQHdCGFz67qUPTEg4Wl/60WIWIKrkddxrWTiR2qFMu
/C7nRSjZctvt/YN+ImOPC0Kapo4nBO5PkeVZIfKWGNqfPX4KV+Fo/kxb2kfMIgjz1Grg9Y/hAyqVWAgyDaWug8DV/Rmv/DqLj67u
Tj+9uywKpRGiH689rqyZfQKvY88dJu/5V7mZOO7cHeO/HMGHel3AutrUwQTNfi9quIcpyrjhF4OjvjbKODsMhv50RCtmFAztMDXT
4bsSOw4LOuSJbQlnwXbbDXEfpcpyFf4xjK0dJ3D8WaJvR+wVuO7XKDUC0Z6mrSqCsGvLPjOEa1ds9MSGPiByh1CDtnc4Ilv1/YnR
TU3VM2xOTEqTTcamwqGKGInz8W11Y0TO6Ml+KAHTFlVb0UOnwHh6c1JfnYQEjJ8WwL/SBfEDX6rquus6YzLxRYswHvnqEHA8xf1P
yRZg1xGfVH+QzsCNaFK2o+j9lER4I1pyn572rmwR5Aqzl3R/vq6xHRKpKGz6GjArtNtKNH0RWUz+d0+csVeNFws5PbQqLnSCY313
/8cGgavKcycjAZuQyiBWvFMX7OVAOZStB4E6Ux4AvhtUXeXg9RRWVbMJQvOOhK+fEYVF1TlcsHlrgJ5QH68KqDcuS5qM2CHITXcr
NkjPNaifR2cZNp506U28A79F0JTeykLEybAm4JY3JisSiRcAcPHcK5tAQb+YIXIBMCYcTQ1aD7MV/1yV4PlXyl/Zxu+Avzdr5/cW
+t1hcs5hEVvLJ/6nWa0y0jwGBP2yjvp6fAN9AxZ5YVELHqryIEAfADkHNUovsb1Hx/GShrfnc05lW8s+3qojb8Iacie5l+Vi3g6F
C2n6zktHFSijJyjcolSwnX7/6NFzvBV/wrAqBxF8LGWKszRN4wXGMaY+/nPmVAcvLSs4Wl0UX4UoXkBuKEbr2AHa/mR8lnJZKBY/
GsySI7xIjuPi54akDZcJxEK1gbX98u3rAu9Q6TRTU0s7DolO3Uu4sNF5oF9yeK1IzT2UYsRQ1iblLznrEBjvEmkSekqItQP5zdFD
8U5O/kuL6oVO6+pVBh8WSkxQnRj+5vR6rRiojyTYlvoB/i5QVN1520aZ3dm6g0r7dGFwLJTjLhRxKq8cfJxadyf+Dk0EPRmltwWK
wCenNdLshCNcpHniP15gL4jhyNOt8OBOUL842Ob/mWSASotW22JFpbCmgZaGBX4hxW4v8O3lmPQXmUpXh33Ti9nb/NYBDQ+G9JbO
LsWBcZb3VP6Y7uzwq/d1HCx6EGBMGQd8h50f6aQUSe6pypu78AZYT5DlwyPJgTBzhOXlQXFagZeWjGP9BIBRr1WwPVADHCvkQZwQ
GgpqPPV3m37YrXrO4Yfjr9Es3ppqGkwRbVM5UJQ/Wc6fIGHyBL6TX3fXnu7Eb51IG7spoT3qRXlKWl4tSLii2AMQOKOB/bQBnnD2
hicS/pHyEC4bo+ByTXFdYdZmr16M/A8REAmgxl5WMChtjFLAWl6wRivvJ+SKxZNN5Z9QUNsd5044Xlkfy5zzaHKG3vwYGlOYx5hn
ZrqqlqxDIna3Oh/2XWU9tvDY9crKw+8xz+LHOdBc6VP4g36IIAmXx3pLQ83RnJUsxJ4MpOlL2892uRDgL70D2+mpTgDzIR7dP5gl
HrceWAPn/APmeO3AukjqCU0dRCuxU3eQYwQKR1B00PVyJQetUaNDuDiM/2OMUGpKZCRpHi4I0LMA4mVJWUqYR3iUi9p/FSN2/dzG
LCGBpiVNyuzEG8y13uhzppqDOVTkSWnf5vbioz8N0c05QbDCKxrxx8O6OWKnqs4SC7ppZolssYDGxqf/kweDK3EhzN+WJXlY5kcf
Cze1UDzIzwIUo8YTS+RIVmQJ56wBLE9tONo5lQTCjgV0R2eXac0sKslnsG0kK/NTVq0Y7IQGfoXtRFVH6O9LbKOglbdakG78lWKQ
+Xxk2Upl9VkJ/OhAvQGOy+wUm0p59/Hasy25ywDHWjXg2C8t3UDaVK4+DxiHAaPMisLZRDcCFl/iuL0tI/87SpLWlF5ONPAUlGx1
x8DyDNzyYwtA/D2xZYaYAVKWZeWijWVbfqS3sXgWczZPctAbrrQ2Y/5xX3UX1iZboUW21jNxFC0YM1uhEH2aLuEB2qwvvm1y8fEs
MK+oW0t4emDC2Sh0NLfsM3RJbfYa8Ji3wkYwENHkBSJ7UXDCATRVfD5oDhSyfIhZa5KijrUP9RkQ8ACahL9V+AWnZQQWWz+3XtLr
QgDfA7eZxGDS/WXysJhOlXIdXFQwdg0zOoAcTGLyQQ6XE1AQClbL9/URmL78qOVmwxBBBO5zWp/q1wb0pLFVzizeFt4OLZe2fTgW
OMffwL+bFPYxidnctPZ01rP64p1ZbebUzZ6QghT9ypHl2pDWuVx2rAeX/Bldcc+NscR6i265VQGeNc+n2HEXkW3ShpPAGfpEHZsO
YgVvmTkdMRDzCZ2kba3mf4ZioWJBZ7SEFCNzwyLCCCZk9QDdBpZoQN1e+sR7rBXiARyglmefs6zZ3s6KO9SoEnhXI7PjeHs1lFd3
4VjnXckjVlLlSC9eeBzUIEQ8Ayu9i1Zc7G8cFZy36E0TUOIjBoaFZw0kQ0cYc+fl6Dhx21F0UgzIV/GsYAZ81Z6NSiGhESitOwPK
rHfKaGwnzi285dQzGz3xVOH8Kfh1WX5aQTYbGSFzD43AeG+AT9N+bJTiYnS7lhAt1VCzhTLFqjaqQvAjxaRhPd+6Q0C/a2nkreiI
SxH5U6Gx4nZm32I2fCIQP15q8Ni9CTF3f20cROq62ZudtL6gDz82bAe5Ok3ci2j8C7PBMSe0Zkyfkc6eq+KE6rnkRL6xjx1FT0A6
VcKPg4SW5m4iDm8hLC354gZ0mrHwe9apNsLK6AaVRKREF1oEycdwPxj3RBiG1aBc307ydcaQU9vUHFn2jI5Wpish5McnTVyyqXaq
cwKSBSSyzG6+OP7O8AH1Pqf9FboroLbGRnbNgfWz2Ci1Jl3HFYp0lRcpVU5HGaV2cwjiqpGjI6V8I3q35vhb4kLlH9poT8GL1Ifv
sQbtw9DWEkCWj63NBWpULMYFq26EGy1vUFjIhF+n7pb7EF8btr7P63Z73ZP8R7nP9PL4VNXrq4RAEg3Pfe9SukaLrKIfIa/dsftN
LAB15f+cQ/eYRfauTdvWtptKnqxSOmFMMX5p94rB5KugYRvlB080CDUeZklAa5FfIV+hJ60+8nUvwoGfeJXSdADwtKQs+k5cLiZW
ziFriVz1X1C4KqnUujfBwKUr6B8GrDOqnBCb8BWY9JJksQifYSj3SsaKhjqkud448um/AZzuvsN6M07eK4t8PhryOM22wLs3hMX5
9k/K6e7vKvWqVSE/WVrDspjbiRBbB05A9fZc8epfIw8fkCBYcZb4dWpC7ZwaxV9H/+baDn1MhrMqTfCDP6dHz6bH1zQ4n1QCwDkk
VKR289+cSGsSqOAIyA6xjquw/a8DJTxutelwizGtcWPAZATOYeQb0ZvPvyPXnoS5X8kZE72yFTaA3x4BK5STKA7K8o8WgiBegAhM
bJ1J1uoZFRwxagKnkxAbV9VeUYmzHXo0E2txHgDBY7HqRc4BIZJv0mcP3z93GFE0PJUKAr02GJSJCgxxBgsatHL+myolhkXuhkVP
Erp6ec8Hg9rRTD3jiIMDUSzvNPkza959YGBXRsSZOY7XC+iV3+BDrvHrMI532hy46v6J5M4xTrEs33H619ioQlF4ydibgAy8UZ1R
C62Nz1YiCQi58Gi61SvG1KAQfVitBL6oz9ogQQdRb6qzTwBASKmQZi+LJ5PlQH+nhzIDIriQIfMtsXDHBiAx58vhxvfrrVSLdA3j
gjUloMZcXvZP5OBUyogWneVkPOoQE9S+ezp8WnDmcr6I7D7Pqre0fqudcNzUspJa8u/Tao0r5JWV2a0wQZ6bHsJUMnbKN6jC7W7H
99LJVWqlAGeW0PxvIpex3ehTzTpxwXHDPHACTOQVsaLiQTiAgEt0hXon8jkzZWmzT/FvMtqNgunZgTDBZPJNu8FNANVDDi02K1ud
ef2RZsiYhKknOCMs5bwm3Tnt3/RIzgXD6oMdmmiamNIygRQ1eMuuC55DvEh/jzJMLZ/mRNQi/FJPfCTv7unGKoViw1dPXyCe17wp
tktcQyx69WgdaT/q0nLcLg2OavHmhg8qFblZZSZB3RRkQcok7FPwqNBExe3OTjA1grgPDFyUHd6/V89130WTpMhLqklxZTCiWXEJ
PSQKLk/F0Axv0en0ZHrI7ogRyFxkC4WtlBGycNfpQplNl+RY8SKnoEvrOCIKl/BAnDzrJjm/55R6DliOGvjyS0UjDzRzPqS4km1L
EfVLqqwItsDaM9QBNZzqnLJgVouEjF2ZA/2mjXfjkXF80M31a5xkcspFmTJY0qcHkDqhsfE8/2ovuMrRChXm6uO69dzYyMV0dwHt
ek6yR2yTtRGP6WZfnPyh50vCVGgkKTFhcY5gy3rXcfuJECI1stZspi2Rr1BnltzagEPO6qmTXJTkMY64lst433darkva/gDljDIp
CKhTw9rzkmboJ+Rj6skKGY0Z8+ahKAuKWiLpcxt2U7f2caz0ou2nxxZIb8VtxlYquZRspL0RtQxdeEALyJYpBMz5Ym1IknXxw7mq
z35x7EAy45fqbh+E08B2Y7a2OpxPh2J/bXKh1sNtOOs4VdsFthHQGGXTJouwd1FXy4KmEYuIpbMfgPyw/EbStzW1TWEIrXUuO/gl
qUInkjpYUu21bahJzfqgQEx3Sh51g2CMxiQ2WqIUjv+KZSU1iSI5QDt+VFd4sJljl3GZv/WicaxCaSWB9zMO/jA0rRS0W8Dmx4p1
qDCYAGWKSMmrrWW/PNWUyZA19ha0qPETiIomrGUz2nqNf8mPmy4fVukfRWpW6MIuVQT1RSzO3diM3/I7NGh7GBz3KKp9dT74s88I
Z6IAkWjZzsOtvg7iAYe/aMwsgwEjVm1JWLu/akqqE0efJCxWcjOlAdjnB3oNMeMPRRgrd8gg4sdLSXXUeInfNVCUFSE6s/tTqHZV
X5gT/YlMWHzPoWcIE6zFxKAE3fBwRammvRKRDzymgrYRi52fOvXLrmCeTlFWnJLPgIGw0iXK8pSqPZ5L8uT0Q1Jr8AcOhp7f22ha
TblMS3jo5mYFyvxJ/hs1lLJEpKdm3KqZjcWS1wx7J1g0zvaYGfhiElH/mT2JVzKSvVTGvFYRsXBWfHwkRRwf83OpaAFleV7g68xB
4v7IPKk5bCj+aqhCl/VbMOfqsGR4Ek8tsOTlrNuF7JHwBMgAE7hOv7W1lnk7Jinuhr6tQHubZdvjVagGFRF/zIAj3IWb8fDxW65x
PQvtqXzR/DPsVdx53IqMd4IAe6uyfCoAS87/Xz+g6YTN0hGGm5nCO2O02aBstU48MvnZl3HxKtjz3b/vgOvDP69L2GAKjMiHvowU
ZxE5K97HaW5tQPc2gUzuiMRWz75KN/ZKJI/Sh/ZAKbUyTPZ41zXmAS7HmtsVWgQB5msFg42n7Vq3zLUZenaS1lcWPv1f/qcS2woO
s/5lmdS6D5fOr+uPKQYGntirht83ljzlA+Bo8ivs/Q5reQwlU1b0zexuLrqFwTUG2B3zflvTBVcgcSzT39gvFDcIEp255318d3/A
XNwVQYMFsOQeeYTUWqlve/HFrG8ULbr34kL8F2goAKNa+kBfQXVD0ZYwFREpXcCMZhIVQ9JKK6nswr6R6Yqpj7z4aGPyJLg/g9q4
v4XL4Vpo2regPf+fDdWgWOJX9AfSQtMl1VwdSo2D6Ba6vDDV6y5ezx10ZR2I795TzBQ1e9mbFDxrILuMaRwI4BtoBVroTCUDPk6h
N5DUbxDC+e64p8b5Le8sEMHiBzV5wcf9pN2+uunBV47EXNSJ/1c3KuHnBQjNIlclc7mmX8GkPQuv5QtlG9Da7rq0wqGBKFcBII/e
fzm0oJoHNTRbS/GUDCUdjC+sbwggDD9Xzu/CJk0evIcwFyTyUXynxIicOvQ8ueXqe0719CYG3Pm/+lEHrNVo8EcGDzyQ3PL10g0z
rKT17SC5F5YVa882mc4kgAJ1SC0yKMGpSeds75EvPH74c2mHmllLAVFiYooAhWNpKFoZbtTXtvJLIPN+UHOhPuuVpUlC3bEoUU3n
1UVJi4BWB6iCT36KLR5mr3YwCiBUh/r/bO8Rf1utKJM/gvQ0An3aN+vJSjW4SgYuj1iXKnN81KH4jaVYnBRkTsjDfDuHLXeHmJr1
o6HpVYO0diIBhNhqt5ezbso5Ee5wA+W/YQuvV7WfG0L14x7+o7LRivNXZ9uNlmOSd5HiYA6Ko658ndL2KGCvfA4fqbLPF/95Sjww
q2SV+lfDxqFPiBZb4zvzTlAh7enWhOpK6ip1DECEv0W6WooIID9cFY1iuJLa1yr/ebbUjFFZdhwtqfORBmXkC2N5OJplbJ4Y09Nv
NK3+I1nGfChHWByCH/OiOG0KEQt5UE3pkvBas7pdy1ye9WVPP3Vd7uRnu4Tuc8XYhjPsx6tCmeOBTGd8qXhJ1JL1MWN3cuQDpQgN
iNZneB0G0X9kVkxZOH5cyoSJo7dzckA/jPWDLhRVjM969/fOfc9j4qoJsoNEbVvxHx3PEGyGsHxr+ASmTVG7ghj+3Lp6egOaQ1d0
Vz84TM9R1OGrde8DVZqRm3pMzwIPAzMm9lZ+vXjNZ+GKr5/1sRchKLJkzzJ6JXJgusnYE8Eg9NFBN8BE8gonOIpTCfa4wnlXcdcS
eQmHgzAQ4jwdKkvMb+1RFTYnKCL7V4Z+2Tg3Z3iNM2yjkaYFWvxLKjegORCm3Ah8rxjIMZ1LKDOpio4np4HYnC5HosffcPb/fYO0
r92ULmlCPgcHtgocA3JxcyrtuLNG/xLnjXSRTHwO8O6uiSMcbEqIb8bbRiWDRtOefi3W0n5I5nOji/6XJGyWE8gjIOF/NcKGQ0T6
RHJd462p0Ezsid/qriCpKI4x5rZ+R7uBgY+IXaqAxUBsQuUC0s+3jQoWybEhyXZOs7dwDXOT5bqnUOIxASpqgMDssmyPdEntDj+B
j6VbJpGuSg6z+ySL5a9DHWJDUCqRQSfEFETvob8OBmBAIpaWiul+CNT4yFp45CZALLSuhB1ho7LZl9cuwn65fJ7WhoEj5V9KdFNV
JaLZ2uWwACW4HVkiDJ/Em+HSAVV4o1gyEYzCnu5SEjZrXoCwOQnqobIdcSpcMZ1r62s1BwTDliDT7MQE+SoO/McWu9O8gauNQVQW
fanvE1j4kEO9EanHWSeXZDwkuhrZQXPRIzVZ6dF/os9S/heWhNg4HDav+E6rchvD0Cd0myioqhC1g30SD5IMVnz2dob32Xln3iEQ
uuGr+vHpmp5pInfKLY9/stVmCFmKpci8nSiyNB2iKIS6C5ySe5r5yoz0Yka/+Bv8+ToG+fnJQSxFhWHBbp6as2MiyTkvZps6h6Tb
fymtw5d9fnvHW8Bxte2bfIDJJBU774dAesnGt3hdqm3epZTl9er+A/iNiZPCI9+gTR7HHLw1uPOoE+6uuRf5K2Lt8TUvubZNWmfY
ATOzEnvDuas5XFj8+Fv0IXOckSDpKuBMNyPJf/Mc9aiqVlIQXwLBKHnrvhXjsEQPCBmDiPh/i85f6kQZydYD09sCroXPgngvOWGY
nZhxH+hvq4ezkbhJtx8KWco/swAUoLq1c+sG3rm7V6EdYGmQKWvURUdnVP6E4rpM+awXzPyeXUkuVOsbZCM2jjCu8I9Aqqq5u7fN
xqLQyWCNU19Xqo48J/+yRWRCRoYMNNC3mzt5hc7N0Uz5o7+CMuCiRoOG8CdHPGTufTv/MMqFlYu2+ogkGcmTkenAjnuW2JwDKPxY
Wm1d4nla5k8Gig0uBA0A4Ba/ob5tJagifJ1klsl4vP4vm2a3wtoYzRxLPzij+gdz45pXK89r0uuaQssdoUF5hDn3LNhhcla2W6E4
2ao6r0GwpybnNMVNKQY3Tf8SDuzHMZvBHCNx63+0xykejo+1mRDqhC8blP2eIDGIWo8/Bjmr7vrsci/05yFAZDUe7k6P1yPFHESF
1Nywqg+zZchtR2RF+Zxmxh5D0jbX0WeSqiEO1gr8YWZ1xLvUTOIeXuRcbWyheNT7cSAJbT1eRFBH55QjBlcfF3wV9Chav5LkXRpo
UghjxUwv/Wd2C6/LZQZrJPZX7htIMEeRl9KtpY0OTgF2gxqkU3HzQW91iWnHMoEkGCoiJgghU6rmMY5T+RU0g82XEOQ8ZNTCe8AO
P9bew5Ukqav0M5XrQO96GuJw/30OYY+HkcE13cbx6mTS8sabZlK0E6KRCcT6AahC2CVSPua8LPFCz4D9sDdk7chn2VOvUtTL4ZqK
eqXo9Am7YGMaquo1B84Z5JhzDLmik1kMFQumqMDEHowU9IytqeLC5ld3yJ6I0Io5zIYgT4KIFcqb9AwIBX4qj3u8HAl4iUy+AZJ8
G9kfX2PgvaNULReq4D3reQ59CcVu2VypimXu7VvY0F/WLz85XWKbB4NU0gPndJhQCsdUvGwgS9w0qTY9rc29l3aij2NeN/lB2Vxc
68W6tMsKWsvKHy7OjNQB2Adih+BVuAroM4BgSsjMy/BM60C39B4cZYcrffbyv9Bv79aRxYVU7ebd2pjBX3eAwg/J2yKujNXxkH7R
UcFdWjIuB6IevuazRjwRo/SANJM+s1+QzmKb9TaS/WiCr3PNNEGbIC12yaugTJXVEUMsqvv1L7rA99zyMWAc8Xod6BYQzXUPkFOa
bWYqynPhIpmKmRZtXBk8bf0rPrVi2BaNxyPuVx9yWhGrEe/ncq1Zx+rICipHxt59/qk5Rn+OGF6xsbMvjpmkp/xTFm/M/v0Xk2Lq
YI2jnq/cto8FvCyQt/1ajJVXooNmBBXubdw+As0IR2OO63YQj7S8dndP3owDnDjxfx+4+O8Gn9M9xDH5w7EwDc/ByRt8dRdOyEXc
tQjgwIVGGObBMkid7tAU79RU4Onzg6ZszxokCWlvrgKl2FzjRKzg55Op8HmWGdrTJZpxrI1jfU5BwyePCEZPDhBppc2GYoLP6yVn
RVYqE0QKIpP3clBM3vvFG2pkWOC4peOOwH2Hf3jUZa82cOWDv8vjE4SEZREh5Zm0Px117YXCqVTY7zl9LVtPKBkiGK+C/nH4ywuW
WSV7Bo/quAk/41p72DYDnnFr6cUq61+EMSGx/CxeeszDbN/rmcDy6jWh2EqzwmWvBLCY2PUCy6uzbqgc8rA041aqT8uBTzdVjsKP
t10cTVTUco7rmmKe21IwhliJuQMOxFWXtdb56nmsYeuagieVhM0h4dhg8qShXnJwG9M1ZLfsVKSUkw5Q8sa/8ZGmdtxHk3vLlaXe
IEff/krbHN1J1x7/LPcLywNxGupWEr13s54sIYkWtommX+265rr/fcNW5GINt4dIpUwGSZ/FPqRjFtSsuw3xO36Hq24Wekc9dhcJ
SIQvTksVXjPmPK8+U0+aEudeB0WmM8lu3qmLMvIWZqjQTwkxXIZkXCNvYHvCuv5i592FhkXWduby5ZJxPkdUshMVYRPyfe2F7jS/
6KM2xiuFO3GWofQXFfpAZUeoj32GwCVNnphbqWePR++LL34gliAb8GGuFC4Sh4PmguMWickrUs8TbhQ7X33vkXb1cI+FWGbxbQs6
uzfYKI0D/frLWUxR0TjiAk11lFSDTNcgHfPlcPeUkxypwHhFKeF9I1lOn5RkzJ6I8CHganVhulKjIx61UdCzUwVrmwG43f/YQWUd
2cQmSISUvmQynzZMx6lw0J01pltZZrvOcophVURnUzIlILBJe9EwWvC/PWe6gN72WEoEqhTEmqvZrMBx+fYlNehlxvTXNgCY9MTd
6mrR2LgnZBwWnRq7HxOu2TPW7cYdQqBpcO3OMBNvAz4dviCVVRY65m0w/FMj6XgAAhlOV/gWpdsC34umUB9sOn3LcBnZPskg88XO
EGvk3haihocmJv3re61ZHdgGKlK6xSZBJ3jgF+IUzZ8g6yP+MDhsqc2QQv+rI1x1wlRyW4oVAHcy/IkwIfHhlthLrl2eDpxHx3KX
czsn6u6sY62R7T8Fm5PGYAAV4Y+Wx4O8omfOIVPDUjOcx/B83tu94ppGVGvY9o8FmNBb3+MOFxo/85pmbk9MekzW3Iwa/JQiKl3f
NuB7VYdttsKW7KXlZLarqcQvCtwnSwmZbGWADEtDqsoW0z12XTyq5OikD2d8YedGUCcCKD0rsAXljeBIEO0Lpz2X5/4UiDDYPj9Z
FrSOKBmI1tq0A1yAWygOly8Lt5wD6JgWzuelHNYADTFE98oh+a+cx3d+tYZlXJmuAxTGTShWWBrBg9T1xmpNBjlzB4yfkJTs4THP
SY+JFxuHAe6SU+7YIgLnLV20rMMi93o8lbokq2O0gLpylqOQUk0m/CGl22eFmkkLKOswDD/lfHVrAJDDudIBVzMavxla9EqzHXB9
LRSYr3uucDL2KGsYc2Dc8u55+JETBUjKtF5csHWVuNuAtZwe6P/U7VG4Ht/y/GVjIVELgIrnNDGdzvfLVS7WsFVKveEpdkOydQgJ
YY8cZQU1U/1Yntpxk6HL9QF/CepPrnqNN1d9FIHB8SkhEKGhvol0gccU/RHQdYGuccciy2paSJLwGDFLfhxZdWwi+VIcmA63r1l/
6JY1RS0WqCsw2SglHN6JFdniqF83dmGpH+FD7Q91v20kmDoX/VvasH9WsvdcAzUXnA0N/q8IrUGR4cwhBCSWo8deOBfRWXuTW8cy
6f75CWJpCq5elDGFl2ITit5XYbxb6zJi9g0uYOX+4ckOP2jrscqKGXez3hXfMPckcaF/QyPk6M8QT/JJYoI7hvc3O9PNutmK7whD
q57l7GDDn1uSLlOcr44d6M6EzqTNb/UsJiBba1B5jxghy6HnD91eK8YJzWSvE8S3MDM5EaQQmrPipAoF75cdgnU6JM2HSdj8SStu
93nL16IJO6t5GIs6BcyGKPTZOwO46oDSiIoGlDZdUUp41wFCoG/3bJ8ZmnqwSdD257U0M+2y/Lc5vRpMVpXpmwBTyZS+ys9WFS5t
G/ikGoFgz1ME0ElzYURbBvIb+Cunb2IxS5b3e/mbJmXW4GDb3dZd8IVSYLzDUZpl/3m++MfmvQepGulJW1+QgDW48OdBpa+BOowu
BUv0Aa1LU2hAVf+wSw0TwPt2uvSeFenPpFRg9MuWg92iZ7bxNWUigETikpj3wMgn0myql8ny1wf7Xp27mKkdElSVoi9V/XRv4cW6
C4iMYRX+b+B/mVCG4bY8cL2gpta0LwnEBOyn4qJXa+aNB4PBqAmDEca2hMVBLnkPZYqbB0d3ax2NVuOhuuMtHGSh71jOMofIi2sF
eAKCqg1am3k+TU6U9hoToi1bbDhGw9PlkHdyhWeyyl2m+/lkiupbr07AIQMbHpM9KxwtWaRTkU0O3tcO9jA2MwmRDuKOOocI326K
xjzd/9Juum2qxdCmlW/XBO5MZbZT5uHs9ihBpn4vnB+fg1RGp841a3Fopv3we6x/4glvFTT52dkGxVd0dAn1rcdhKeRV/pUp25vI
spPqMzaWRL9S9+XwmVkj6AJfwevfgWs35NJ3bb1MxqaHx02IYAwyhHabEbaXeHOk0/1AwfXw0nYE51GfrI5Ca0G9UDYuUN7bG1Lr
UREeNLbyWC0HYivTAmkjsSqIf5AwdGC/t/wuTzeOqrxg5PTVXdrYnQbMWoqn/Omvdf7XYY5gm64utGwSlWynRo6f/d8E5tNkprm6
7oKF4ZqH0LS3fVW6qJksNLxdWkV6y/wObVLI0KHADegylATHsL7NCM1PFbRPUUKc3YNohdaFQ96q07vhAXXdtcfyALukzzGFAPmV
hGhrj2a35sTiRhERLwUtl3B0INiyj1r972r8QEzzU1JK163cnWntIMytRFjSth8HN2U67C6/IDCOZ4swVPfEWH9ZdlHS5ka5b/I/
VMoT75A0YWyrvxHyrIj+RDs0fmYRorZo9ulZoY7sVZPHLSFnCJVGpY00x//nx3g5nn4RZA98Jtgd4hG9Nq5UiF1kR8xyPa4IInq6
nfIGJOVAu1JQfrEL/Ze8ia/dyKoamVK0f86joU3MIprWe4HyfA0efdw8VHEeye6n5Xp1cUN04N7f9chSp4DkQiqHsh4QBX/tjfR4
/28bfzMi6l7FHLW7QN0JCazNaxY+Y3xe21O1WcBZ+8dVa2In73Pla/ojAUs4i3SwZZ92mQR96ADotoJDMiR2/fSCxC2eErOsdnns
ojnapnJD9pVxXDxX04TGIaEUqCamJNmwgY9ShTNNbCL70RGfQsimvnBqYPwZyZfLuidrOfE5V5xUqDRS+AJg/gj8N6+dx4vtftPA
HWgr2BH81VPZQnwFCQ4WHeSzhf3SWj4HHfNxaYMYVE5qC05EMuk6g55GByAs6op/KzXAuLHWP6WkDC/03IvZssaHI/6+4NRna/25
girCnBXj1JG9Z6xNknzkMgibDJcGePlGQhcEKPzWnM4x4bJYRnmQGIwJwLuiv1VFiFOkAjKFDuDogewfx2hnLnQE8g6Kb6F24+Sv
jzv0C9MDICDXg/hL1pvAsWD1xfhYNbJK0W+JpZtIlZ7RFBZBD/OT/nM459+hyvz3kKnIWUi2Amr/1uePa7ALbGMdWey5wLDGOXox
/MNEOFv9ue/rLQ09CHyN7e6wyRXyUsG9rZuGbf+7i0pumYShxERKM1K09Yuz0pNf0yklIN03YvIZzK8ImEdZdrVE6zrNtcxQOHjL
40GNu7mJkF3/OdY+kjdXXTGaJUNSwQa2ThrKmmVa6FKAb3J7Ie9QARdBiTm5XxEuYEgzu8AH85g4qB2kfYfHq94ckva8siBFvHi0
zhEwn40LtBGi9IO+PAZdoFpDNPys1ybj9vnvbHFDyRK7hZOZe+B+i+cOv1v9RtBwxHDYxAMw8Mg6/WFRzmYw9cGMaQEqktyJ6Kg4
NyXUhT82q1aAEg0TIPcBcbigkxYoPXN5/ecBlcXFOkJYVzvi37F//7GhvMfDzZbOt58/Z0VFYwE1q9SD10xssmvy1o4zwJH/0qIw
R4WNb8CrHfvCXgsOf56EH/lQWeHkucgmnHa6CbVZRkmNnVBUdnk6GJxsC1uiq6m6ro1HAiL047Czg1zgdzwDfV6WldpkeOUwKmWW
PDjMEGdVpJis+ryAt3j6qEP6wLgryd6N3TI6UcAoo8i3kRTmxwVOSFm0Hp+isIYEqBqi8Y44DiMBql3Q25gB0G/MpYsao3EKLJl6
F/zs0a09TkkInLYhR18WlnspTntHJLybQWDVn5Czupg6jIcCu+++oko7ZdDQvHrXY8E7MQlZcSl2PJJKds/hY1ZvnYHPrP5+QdOq
CAeTIWd7LbZ5lOdn1vFk5+iQ1MR0hxKFQSLCOHkkoN8wu9brzExLmnVPrctlWtMdf1p55jwDp9TnsklxTRBRUf0kfHrgD1NC4ixA
SCXJYMRLqB9ZKKDAPRXGO6IbxgghbuNC8MPk9G9/gGgWAAWIe61QyYGR70/4vgV8L7Wyr2oRhizHLcQZvxFN+4VhQOGi6sT/lPV+
N9nGYMl+kPutjQTqb4JBiZ5vQcTkYl+lPxVJW0FAUGbxXzgvN1xH4B0CBABV4rEfwvjw2gt+fgHiAVpvl0sVvNMJgagx744Nmxda
TEmXfCrKgxrWSm6g5Ku+d2044bt9ZAHCYmmj6ynBVjtUYN9ldhPcfysxbBBmpI0f0xmUR233Ad5vXe7jn5USu5s26f/L1hP5QZUl
7MD+fM+WtIrKi8sXCPQ8J+m0/+7zrNx0ApS4JRxLw1G52jCeIcMBZppEFlwc40Ck+X5vLAN+jeXo5OR7GvsY+TMwWGWoHqK/IGW2
e9Hns/nl8Ke6XFNge4uCPP9OGbhu3C7KAjvekbLBmIJ9csO7weRzNfWrf/ashSAq/AhPwvayT+M1mOtYLG2twyPQh8y14Djnbhlu
+FChK8jubweF1SL43nLaR9vzhq92/bvuP4eW2NHPc3kJECYv9VHyngO1t+FDSIsVpjzBkIX/cIPe9tHYQvkAN/n/TybWABe9nUYM
cDa7MZWPYnOqXxMBh/Lr2rpvTd8ml/yeSeO/4O83gkOBN91PgB5vHAs62sdXeSgB6Tl98g2fIiYvkmo0b6yPBzXqGopj1JFEopxw
Agg6w0DvTNtNpQ31/hqPounqYGQbsVsH9y9sh9y7dcbFcWMBzyrfHJnyRsA4ZMZP0jigDnRS4mh5ZTFzRAckaa5cDRbtmMB+bIfF
XdQqi9AcBd0p7UAtyT/K7DcbXoTXrjKO1rpCagzf0kazU8xdeKuiW0XXS5AN2FKN/byoKUeHRMC+eGFA+rYoJY8D5EpBy0zSmljb
0ltvej7nks1sg3+RH2Prjvs+WF3x1rrCjA/9DnoKjtOLddE1C5pvIb03+p4bR756WbEbkHk/r5jC3qx5rTYcBtpVU2V+N+/dZF/7
s6RxzHN3w9q+dGL7iceSDoCA5i2lvTUhpyH39evKF28EAbVOj7d1z0As9T0a0+IWp19bPrVdNei8+7FpgbJMfySquPGlnRz+2Km/
pSO9g+zWhXh95RlZVnR67NSAamiSycBd3hCvwCCmd/39PZaqg+vVcKWhzgYx5QaecGbgpmeckobHL7Lz2VIIBAwTab7cD0UWFtwr
imkGXB2f12I2Kh5EI6pA5DLfvB8B7YqNX7gj6SPKJu7siQB6vNFLjrwkUGo9MDMRyYo8BQ/8K6Rnipq6OQ7F2ZYt39O3qjyDQbbE
wxMlhLZz3Yyz/SpB28DF+vRW9f1Sj0lVUVRt1uyovetbeJzFaWWuEcAArvyJVBviFrVVPlcMH8ZG2RB3lf66iKY7i1AQY2lQNJOp
4BBBrXsJi3a6Je3ZH+mY3jpewCzcif3WhBm2Xp+lqSYeWfSQODXhvpkbno/9g+cxWnXvAUtQLLxY0apIVNaeDl14euVc+ilQx9bD
B2ybHIKCVWQ8EE9rXcL/YwfE+zD3KGBK7OQSF8+Vtnue6raWDZej2Lu2ff2zYFzTxJnyauLL9mwo//YHNbKLflKT1TcTSc3m8wyS
cGIMI47xwvvzZPGVPoTfRYtNyyMYolVTZuYfsJjCjm8kQi1jA+zyP7+cSuyavV0BBOPOz6vpkjJlA36LxVL19I9GBhFWSBj9ZACH
N17/CKIlSI3p2jIyTy3fz8LlEyUR9D46JaFiEVUahV/YQUae+l7q8PNYJMtvP8pfYs6tAdjiX11aFQ5Lufrl35rZjxClcS3whoHD
TXeAehFvRmb++qhYiGGFdDB5TZIYEHvGXdCoGSXAwIaJFv5MX1ZDRu2lhwoCwMIWDMMN/mIvrBYkysSkPHrbkPdflAxx6720KVVr
jbPkq4WPuF9cJ3WRlPDpRcx9eKLHd24YmCJeCZaqvYeZ1TsUSdCdpNj2eGOEiP4m2bB5u7klygDVFi/qV4rkFXUaxNuPTnusqpd0
nuGXlB/n4cdKnn1WaJtMu763KhbXOedG494dGyJz5iegXgGxtcgOwY550SH/pebHWID/kfGAp9lyWvlavPEXa/OufYvFE9B8EX2E
+y6Qdbg8wj/UqpZoviG3FDJywOFdPzF9kRyi45mPCEJ08a6zmyQBW8QMRw0YOMxvexE0Uu29i2qerJwtteSzlD3OE+hZDSGpUytH
PBFAEPbdIePvMIXDo84evOFI+Epg4PudkS8cAwIW7MuTIcyvWz3J6j32zUwI9/begZBzqarPRHpLWFaWUPumaTW/w70hnYiJr1ig
bmlvLCAsvj0h367XQieZV6lMlKwq5Rbql4wDzH6r0wUfrP3aqDGRPcoIR9vZATXRkvMUN5HEDwOHKqqv34hUBfJ8OTFyLDAeJMas
TRGXIvTYzJWFOWHtp/0m835rrZbmcQvwkSVvjcRtxGkF+xje20nsqlT+TytlsJrQqlhuYx2CqF7ofh33CXiIJ+TABmZVBZ5aKn2p
JmN4pk72eUrxmxnob8Hq/sGeuslRRx6YxSuYetSwXh1/yStbnUfd2V87nLK3RUgdciPxHpZObNNdG2jgdfLJMLx95W7us7Il9hIx
krAJG4E/VDrlJlQhRJ7IhjbGKrmbYYeYD9n9X3QgTgkMmLsalvNeJD279oyg5/CHOiOHzp0SDRKhtszmNYWsapMRhKrdfr1Sqrkh
jtIKPFQcy2Jml/BpcRWpTUqs+sLt5pvUoIp0UwKbTOdpIn+asLdOIXMzJcIK1j6BkHiuXopd3RBeTqkU8j6jobp8qAa7gIg+/lvK
MOl3NUl+nRplT2vdY+87TeEgc2mI1tSCHT8RPNZjMeBB1KD+EaHkmlHolPf1s9xotushwmSBJKDVSsO2+GBR91grP0nehpc01yim
JsiJenCr/a2btw9Re4fNV/AEnmrQy62jXwgRwq0U6fYeZ4vp32fSnwDDjwqJtkoeuC1BeFT6e80RVCF65LryCR9bTi4lNxRSeBBu
6eSfprduafrb1fec94xJlW3a6E2FJf8XdhNxWQjCPSXva6owbtYzNkrmTDNYGbp68ez+r6G9h2qEoZYgDFtOrttevmi6gnk42SdH
HMqTgcVK7a49uLnUOTWi8gDodgnBqmVfQ8EHlAzgsIXt39Nr962qbvBafFnXSz5Fun7oxFkt9501wAuFYAW5RXHcytrJEbKZwnCS
e3+HthQA2JjW2I7fi29hqGefuXamQoJjKkwrCfdP9QvahHFkccQQ8IoHPU9JRkkARLR97XIEtGj3Db3SzqURW8C3dvqJssv083+u
NS7b1w2zthN5fkaEcfKJ7r/gXQ7USWGpywLaq9aMMWThb4Eb/fwWGpT494k8k2f+J1cOLtTr8uZBmvh/G2q4/FioiTt0+mWklOVh
6lynV16hB4TAmFHc8WFb9Q06VTu/lrelbRSfEgqISHGDzQ72QcUHXzk3vilkWP7lB0w6A2wUoYi+GAaWSKRQ5mxyj7ewWL5382Ib
9Dk4BiFtrHJ4J0ycgIv303qiERMPMYnFiAF0oz3Dc10cr6A7LIhBlh5zOx/uamt63su1pN7N/JqRI/VInzk4GLBeByrdaB+ojQYZ
kmfp60NBFOinReIC4JzoR4yl3u65h1Ss7fuL05rWK2+JN5RZCYwIAEtS+EpYqvOWo4aVGHjzRznZEhmm1y+NA5WaNr7tnYVylCNO
VQVoI9sewnFWMXm+0DGpZOlXKYE+C7DNzoG71FRJo3b7ggYPjNwzt268okJ5qA2aKvLiOs5oFHNOlh4pKPY7rJKxSGahfizlkhbc
ZN/SN1Z5/Lf+MsoFbFfnJX2Z238FIOPpmyZo/oFQIdIksk6eV4TGc7Ilrgyfb0vIlzaxz+1Oa5+amrQqgiAqHubkm9Zy0kx5QBBc
bOPqAZ6RLTxHNqQpwHE7SsI44eXt8BfKUZafzouMRNkh0aGeWtkMV7+27bzwPGr6GAdK+s+QixwuWs9nQ1ZU7LOBGXpqhRhyvhEM
yjo1JlJXUmXbhZgZ77ccV3/174QbcsLzf/Xzg7L2ng9JPryoo6oi6Di9y/u3qAby2Et+yumVuC+iQkiRNNsvY2RfCqeeI27zOicN
FLLIvcet/WsijQALMQv3bI4AbHAwJbTaYC3Jd7JTepRlraOMWDqbi0piuyn0pt1EXbpQB3gTTbYpX8BOQMFgX56CP7NRUgWQSy3T
j06vjYp6Du5cuLbVPZMYeZRf/LBKi7bbIS1MXHgzXlgnrc2Cu49CqSpQETwr9CgM+BSApt8AHvOPcABI/sx8qotJPSIiRc6+03zq
tmC7RIm6c6b2QfHEmm7/zkjw3WoLPTcg/l24ykoPpI/PjAkE5UbtgSDjCcop/fEmcbIiq/8XAr/8bZAf9M0UgsBgoFhJSfPrE+dI
4krJOj6LEPMrniQVgcnirIM6+yUSAnlQA8BYPuVQmJULpzZudtd1oapgwg1LoLUNsTCsO2EW2TnORkF6PjHuFfuxe+/E9650q6TE
1hE/9tCtnRQKfHd8BIxjAnpe9Fmt92olcAoqvwqzLXDKrCpf2XlLQ5ISUcV42G3kAZBRDfy/hvRurVbLGwqP2nQjKNDP+ZJhd2Jg
I+r1x55YnpeTf9NuiJ43a5PQs/vxmJF7C9LnPKlyCHGAOFL5yDBLeM2DNwgEBfrn0ooF8J+hWWna65ZcL/dK+CtLHRN5IicwhpQ7
bn1ISLYzzP/BgqIssWP81OvtM8rFMykenVX4SnEZHkyyfTLIcgA3FMrC3jzHO+IvyRTrU/ANNmPeo5ouPTuTP9I4N8Fuji49rJ2v
dvvANnakX0+hLg0I2zc914ePWOfzHDBx2sHPeFrNMfyjLF8cuzIc3dC8u48P1w73SedwnKTxtgkqniuRE+sNLYc3Te2iGUN+yGRD
0RKrIStfqWYxuDWY/aVmqzOzMlVRyXWftmPtV0STu9FV1hLgss5gWB6qcML0ADv6AashtOveQuho1NylF1mdNUQ7HREYCxeZyGkn
cEpc3xI5EP+8ldwnS77qpEL9OOoDRk9IYbFdEf3zEIyaSEsdOmKo+23hQdsgfZn0Pd6M12w0v+FJyDq8FW6hXXg9gjkMKC3y+lw/
/j7x/sYHPsJtGMBgPC7/6rWc3OSS2NOATaBSL3RGa/RVmvvFH8JGvuGZcWbPUmq2CgxD/wvGNqUdH0ouOLWEepgjRZv1CDtLmKjC
htiz9QbQ2IuExClVccdmDzGNlsplmB9vZB+UMlfJHD6e7MKRMwMjpyPfTzkMckcmysac5LvN6FQ8sFqXaCyeMj85RnqWqBMxZMFv
TAs7Jfbj9sSZLFSTYQZfpzeJ4B5cP0lZWNFlrFfRbkHkG8BvIKczGTMy2RcwyayEHYVKEalguRrCUHcw2j9ClW4Pw9ZteHUIZupQ
pObL8fGPYH9/EGR40SEkEH12Z3uaLc03JKv8I4ioo71DGDWvvEhja4ZdGC38vB1O0Qh7vjGJ6YExGpLB8Eyei+YYAqDBnOvQ3ixm
4ixnF+2AnUCrTEkaTmI+u+ec6LSxGwl9pH/eXNBG3QGgOLvEps+oEKMVqsgjtGck87/HWKpa54fPZni1sIjHe+LAT4Lt3V6b04sO
WQc+qSEnfEX4NRr9FwDIe7r9L1NNA81ReHM/PRi4sWckrCKt8S7obUDKZ7FEjp5jvgaNvbHBMYAeym15b5shwKoeHY+/Auhe3Wav
pR9L8Fv+RXIxD3tJqEFJGDw4f3Bp3/lsmv9o4QWfFeqg5612tCIZdVwGioNEPnMo6X2FwW9RHs3VLb6VBLOoFqcO9XL8X0M7NWOU
06SrfDKd3JFpxy7QEMa22AWhBhIO+zgFtLiyXhQ1ycJhjuMMxSOJDWjUWFxgfy9r/En9ihGnIVP6feKvsAyoKMDpsPVw+zUzjekp
c+mf1D+yj93RF9N2Df+OB/qePJE0uDkkJmq/QQH6/y4qVq2p0hcQVQsX7GwX+trhhV5Nlv0Dyh+vRgIC4n3iDTevYiaZVG2OAbmI
5sNtHXd+ki7sSJI7egVGJPS/rv0sCo8p+LBXPP39UWoqoswF9WFJleapoDWmQom4H86cSqfEkGp5I2+lvLSiHiYwZWcuz8MnO0Mk
hzXQ5VBx2o/onEwejDdtTJ0nl/PcdaxJB8QRYPQHLRoR7ehItKGEQv/3wzHOeXLBYJUCPO6w3z/s4VVYjx2apj7t9VfNW56MXh0i
lHvYDIF/XFxpuschjhQnzafs6ZcWgPvPrVut6SZs8wXvtTUPCcG+MPF1/ELmCllaz6whF32ZLuzEEoFr1nfBJWYfELXFKw14cfSp
m4ofoNznPpmvU3DLQcQo7DPgScRIvmTIq43pPKvY/+fXgSKKkOPikbB9mFhihVARvYgfvhxeVir6/K5uozbx8SAwYOkDroEehy8p
3OiYyIs3vON7QJnPCOxgutPC8GT6eBfCZdHkwkPL6Sf072Z7dwnoR7JRqsK83d/s4IBCVaqBkwnZwj4MbqiQdPZdGz3hJ3prX4b4
ePodIM0Fkp1qiqCZkTBl5HOhiJUHcdPmfK+sdqaSpI2sZxC8IQvG0/r1b4b7DRdpAmtboRHi0E+vhvYXwUop9iuDMGEyB9d2L9cz
m0WsfkV2YwkDTgcddhlaJaLVdipFp2rJMSkjeqhmj361RLWYHi4WAj1Zny6mf3rGLvyIYJpWsWkhuw/gI/RK5CYRdqTHeseU33X3
H4Xa67TCqVPKDqjwdJfny/0SAImXfT/KglXyskP7qivZtvoANxC4LkNhO+qa3cCXr8ZgStv3iCXiVD8jxbeNjpr0+ixmY5WM6N4U
l5JY2+6OJDGEdMO96f9l9yhK4gEnnYPJLXDJZRBJlwuE+QTuvE1HWOCv7qMidu2IakobEj7mQ4JujBS7yW86i18/ZiQd/iv6bjIg
2VRznpdQZCwziaqKvAEM+t9Tc5IfSCy/5omFG4wnvvdy1denRTlLyGn9hThvxv2ejQY75rfpPT6fpNr9bj+HgIiLkhspGnZOu8kP
9tcp4FsD/KcFABaXbYwPdHCBIoc2L/1oz1XjVI/FDhPXQI0wVBSu7smx47cPeIwq6PjaMBCRuodVo7DD2GQJlfoCs3pxF8TaF+Y4
up8Mq7HbmseBBASPa8KvNOz91MLOcNu0qIkDnTGL8qN7fSqX5o27FYfEWhKP+a4QhmjVFRsRymuDAtBXIgNu0IZsTLKjvMuJe7fg
58ZOEqF8nKVEI9fk57IPaEnKaL/4Bbi5m/BI1qM6CVspX2F4tXbNlbVEq0EjvXpixJrhD5q4GQ44F/e8pCxYt3oBB7QYHAaOZvaA
AwOU7MIId2oRC8TvEHhiIIF8sT4AFN05ZvhUqjM56AFK9YICY4xiAdYRg26wtGKYIAB3ovAoyBpmak1xb9Bh3liCDRonC+BFukD8
NH9ogzKvOqAUXxEh7Y03U7KV72S9R33/XqCN1yVhJ2WWhAURgl6CVWnjOJeMmYvq44/fwdnwM17fcemc09MNTCOyqlgCW747fGjS
k1vzirXlqPgWE+3/Z1gsoLVi8e/9wzj075HoCAoxc1XiYBwLj4jsCi+nXUwceQr7UhpnkzXK1wX3/x3f01WZtMvjq0uH7PAb7ihR
N2EOrwazZQGO2XusaHcnm5Ienejgs7T6979hs5tZSNNOx+h7n4+ojm6hMf/cvSnW4+p8a57vDzfwax3zVsNxRVgBCiOWLfTsYlip
8L023lzyob7ooghXvx8NWWohBE5e4LEJ8/7W5oA/qZXSMQ1yiySdG2lsvta5MN2MXpnBtuTCCOcT/XiZ+d1gMW8/LgvjOR0oZ0fV
QFWMsfbTrVPGADruMGBID44Vbtm6Fj3Fitcn7QWyeDXU/nbAQWgwzizKNP4qzPXzl9WpeuFICkf24DlyKIOlGjKQ1JeF+7BRKECl
WCxSsKd+KsTkLJkAY6/VIKuKe9tESY+h+qa930DbjzMyjLyz/oU9ykJU1TJ/sh1sUZfYk6Z1vtL0Wk+1IsCdBZqmCXffi5OS42fF
qbjfutrryV77Xho7c9isf3IfPxNLzNZfZVlECOba8WyD77cP+5i6SegGkC109l4Qk8XmllYxQw8cAE+iHm535J0mA26YifwKNnFN
rTmNhvwoBWDgWhpWZvvcqLiY7XstuR4Gx8gemh99ychxqcTc6AgJNnx6r05kiYxl3/ck50TWgBEZOZ+6Bj3lAg/I58Wmws5ZZrHo
NGfS4H6/qyCTu9qRqWgmuG+kXmXdv85Bj+ZN8gKuWbks7V4ibAW06O33M1sSG6VGLxXBTXHmjzAG38WkbIEOTeXUPMOWo1BaAbxN
iIhMK/7JuoLQfMTKP6Bkkyo2TO1CrVCb+5GGVrcY/hgmTNb/VG/W6HmpyFF4k7cFWCWYnUwHiTjlv9z7KlP1bx6291FGGckPlM+t
XbBBQ2n+oTRx5Wx1LhoKe59mZUx+kd5gyjX8IT6rU1vJjwHszw+l0Cg0UpJKTqRIn8urE6MdFuGu0IjdOax1PhyS48GP9xUivh1O
mSMjUyuDV05FdqFh/VcBrqsmzoW/KXyYrRNVA34hvZAnsTbRlTK6DqIWTMGSPB4DnXv5zKYi1byr41/AXUujdv6vCBeZFWLkYK+p
nnYw++NJ+iriGqweqZ3kSXAUNM7d5iZ3Ub1F/gCEEop3hD7VnLmLwXuxvYQKQv5RYRp0l06Yx+wKQHwopg/4Cy17JLEPcSCo0k+9
mXMN+Ucg5AiYotTqnlTmzBl5sNFLAi+PuH4A7sHKczpCfmU9f+y2rFw3dgjfGDflHQoUR4IIGS77hE26P1/IJpJwYMfvjgCxhdMc
RoLPoVd1ibskYuhZzUcCJwPYlIpIO13GLeNqnlS7XRJ3HyZKXD3mBOrxf2p60qz2KAJqlCOQ5H5clqpD0GweLcFdXYu3dhlMj9jK
dYlhrwdRHe4ZB2Mba531gemDdKBVPXTcuEx9qYDoH+dtZ6Pi81WeKgjH1BRzdgcyR1KISnhbhvZhSkUt7BKjB3StmC9ImiqEpkol
qpe7KFlzD2ISENqtXAdF/KCGA35Dmi6kPNMAOzEuO/M2k6chCHHmG5UYmo9h31ML+3AqZuLk5yrZmWbm3Ptg3xDUUyGLVw7OGpr3
i5IUvCaNWv0Jtg9rlrvM6uCu+omQzZlQvFRKBAQLDRtQ0Z/w7I18sZ/53+E4lAbxYkgn8bjYSbv6UUeSIB9bM1R2mXV/gRzmxFcp
rOvKHjWYbomuraswSJ37sCjbtKbuXkZNoGC3XSkIORNVkBggX3DdxChxmg2+m1qfK42qN7aXLxrX7lKi9/3XHTt5mVzejJRP4TqA
huPp1NImLveFq1woRk5s/uXu+1soeJZl9hJnbHmQaUZoCcBwoDUjM+rslbZsTGxRGrpA2rTylF3DH7ffge3uUjx6O4IpahmBvt2X
x1tkA8KqwVzqGfxcJWJkbYk+aHACr8Og3l2ZL16g/Zsjr1lcguROipDhnUxwhUi6vWRw+nxx6g6Qdx3KI49kEVWhjnsjrDFS66/8
CCis+oij9i7s915werHaCH9TEcFhuiZ6s1bT7pC6RZpwBPG8lsQExgGubYvWCiWrX5tT6vvxuzvgVkIJW4kHNn0TfYejo0LqTRoy
0BSL8aIelzx+pfzd8i5Qjn7uD4uoa5f7UVkOq/SiNilX/iK7+tNpVvetuepVuSXqoG5O1RBg7OmmiZDhUJ51ADUqYXRn+dHoncwt
DL/2hQ77qoAWpcKiZqUMInF8taG8IyyxVKQv/F5JEcwIAFtOtPCtP1Ioieqhp+t5fEBrRTAjaHVxQ2FSpWtV2IZAfrVtIL/PlUen
dFhWCANklYkFgm7ukZl0yRUjXFxBLZ78q7n4vmeZ9VWu+JnOdb9cRqaqr/VvNuW9XJeAnFVWiwvZHBS1tUJFuZ9bTsozhMCzpoeJ
qKZ/K19cjq6NiTSfPu2S6Q99G+m+FVgrm7exmrcA9oYHXHd2gO/nsA1LBkTimAqt4KIeDsXqAf7s8hLIaUs/uB9pGLY5Fgr9hqgr
ncpXvpqaWA1aG/63Q74O51A4sV/V7rxqUfag/JC5R46o531q8oZPkQ1Gpdm8SDCrN97Tsk/GkL10F0jTBiv2oxY79BJ8n87GsWtv
juYGdi4K7dI/2F23bYpfBUwb9NouGZ3fM99E4IUKALEcG0+YfwujwpMMpt9mIucvpKuGBxdgioJCowdo6HoSAgMzZgWEuJYkEVO0
nQEaqgA/B3QBprzR+vWSYHkQ7nq2u/p+k8mw7XAIkMXznPkbd2nU2uCOLFHyoXtGhlL481aHUGCG/7Zho7Gb3OGOghy1+5tpxst5
Jydxa7Y11w8RpqFxjf21R2DPET7GoG5C0Yiaj62yWZUxaINGWUrZ2FS6hzLHiNG7FYJrc09d4TsyrxYlykJDS0MLL5MZPwZupkW6
txcyqipsZ7Mac1uNOUEe4uPI9+UOVIUPbkJfhfaickb/RSUGyhzl3NFv6G7r0izihepsqTx4a2saT0Sp3Hg0K9ZSnBMvWpyznD1Z
rVLaqOPezh/6BmrPsXLWq5OazRzWgETFAOU+rAyrmhCI9rqlXS+keffkRTFL0kVo7mKd/JFa48z4TL/zMGv9LR3FVZ5st+rBVm2K
G0RSbXqQCdrnEiJjjYxzBYgEN/bPj5+kAWuXdoDcY8TqpR/F6l0tIePmZllFF37PVdM1zD3PCLsgJ0fYdlZTtH0mOOiim+TlqICM
d2r0P2baLSH+zoT4uNJ5dR0K+zzHc1ncHL5f/OqzBvx3a64vTLTfUem/gbJe7kadTSjWfNnUDcXgRbn7GR6TgPtAjirvmMlSS10k
OabvG6k3E0vnCy6rYYoaOD5alKSDr8WrXZaZbh6HyDYqGTKJn1NqW14RjRkPTBfgL2IJipXdgZMM/deLJ+9D5lLg9kJrHiXLRo/v
W2A0z4Cn6Y9b+UvCVQzF2yDNyM1obQB1xp9+yJm3UrCUoib/87USIy2kmsS2+XItTablK05WGZpcxoX6Fdo2wvD3NrcjdD3CJEu6
IRz406AgVO24a0h+DMmNj/PeEUv0j7j3HRT7p9R/eXOIB2DGMyEHJ8mwOD83PhCg7YridR3lDNO02GhkjldrN1G0qovz6U2Dm0to
0C0zlFL6bc+5ZmokSPDc3MXdXtN17GGMuBqYqm0QqBmIWyZKoWpFypaUKjXNLlBw2DzGxbf07lSeerg0hdUYzkeFEwYLVwWwnsTx
dDxNJunacU3+yARF95tZ9dMR27wDO03SbEQ96f94ShV3ow/X8Cy/JG1wVldL4GtQpjYYaogoDFTSw3VFSskUX8yJ/d6CyFBMEPXb
sANXIjRkomPB+kNX/oGnD0h0WjW4lWoOTmdF89uOa+DAJWfqpPgMpRuRcnAEk8K+LeMcSoljEmbGTy7sEWQyNml/hdrXw7smNHIY
I1gPjfidaDZY8HyNiBvkSXvsGBe4vx2wux3jpiWWeU6DgJZh1tdoHd5wPE7bxKrpyRpvWTUdQPW+d2rpcmV+aC99E3OsYDqkNLKJ
QmxlvA5DVLSDyFiy7fEv0o0NuIws/zI8OBrYUfSOh2ehuNpx2wlEFMv3JOk2qkwl3DJwg9bJR0ZIl1ilFGaxOo4Nw7AWSS2wqY/I
NQiOxFbhQXTQSQKwIc+yR4Dgfju1HeKeueYZ/d47J7B4Cgtv1oF+dDKXBCKGEDxhkAmz+YWxhmdnCskgHhrZwKvtN586W8pHmv69
L08HLoBwp6MYSiRrotbOWqtGZTnx+8J+1U71+HO0O7+1ZHbL6+7tOkhEg9c8lcKb8yABC5bpmnMFZrQd6CdS70+XwBtv4rPYDnPa
wNuPs1R9L1v07Io8RnjuOIIZVuU8OvDUMfYAjxNMxN0NE3E1V5cdt67rTZPqFQ1mr7/7k0OBn9fihWK/I2hNDsP6Trxq1GLVItIZ
1gbcqAIBlrHYaffc8J1EPaogQt1QDNwBLPivGDq5On1GCJEA9tUJ8eQFMyagC6S5o2Fx6F1nsMfFZcSY6NfdFIR6zoCDmvHA/eai
pBAfQ3eOzGKe0txFDz8M9kxJgw6gM8GoDO6k2ZM/sUpTraPOu2cMllR6nMs5f1Kf53fkOUcXAEF4Xft+qcPiVA3y7jWLHZiR7bQm
yYcjYajdUt4H+6n0tDBYl725He5FJrqDy6h390pIme5uAJsC3qOz12KobMu/GChhcLA/E7uHvkymjj105zJwcfEswU2eNmj6gDRb
ACVtwhyWhv5qhpGFg+5IyJ8yxsMicrArXt6REVsOO9ov0JU2BMiNXkvlKdDMox1qFPWCxvRJHpxJt1ubTVfj/kVOrxou6kuCdonM
FNXL6wQlXV58Fb+q3Aink8ncGt5u2TpiPezHs4JxT1vLF3eL2VS3L+z89MBv7qHbQsFnPnrXFGh7KMTSWiOpAoomAAS97DhMtztw
xDTnZXslzG4bLpMBdaFUInq3bGkOjJr1tFTswAmAHOJmOl58H+QjlCKkcSmoNA9y2Zje0KWZYNitv4B96cQHvFEeNoxshZM1rHt+
B3/0W5Ufgr+qYYtUzMcw98Az9Bru1qf4ziPp0a7T/Mt/PZ0crwXwtTdr4qnciznu/AsXmyIqfN+E1dqqfluqPwXBYGNAlI3rsReN
nDB9G0QNanejlIzxkDMhvpEIVUlyHIavSAJQZtijGdPr58trM7V4Q5oe/393O/kq89vny1jWRB8cjtFOUmnvJhzZwTSUwrmElnto
u1XsvGCK+/r9Y5yNQTFM8r4L6zYLBI03nrSBRfOlC+XLplmxAeeypWR5666v2x6NTIY0hIxCXpH1CmBnquNsF4RXHXQ2UEMEQYWX
1loCwqVFZjYMIFS8AhWI5wlRkMDSM1ZeCswyjZlfdJKAzWM1ZwyqvSoZSz5UIKSYZIm3DM9pJGkwxDCObdHSqsbgYE69B4+nFkWa
/puY64GPvQhxKd7gEN42il1uZxFZecADdziRmEr0sOynXDWexzPufinBw0VcdHI7lgtBcMXkjsO0lLPJYcw24fK17PCkQKuNyrDi
I+ZPTndPotGYFxCvSAY8BpR/iQMuRkLpkqgAUCRmWeO7fguI9zyxWUzE9BFhbfdoJqlzLDaEfmuHR4MVEfWR0K7jFqfvTdMtcp0X
03SB0DeEuAzS/wDExmMjjd4I/WBy1t+hNc/fi02E+9haWT9325ZTuGMucbtl1VruL3oWf4lKCujdSW+6eJQfwsK38Qe/0IquUGVw
EB0EI7cz0KXVyqHvkHAJsde/Qn3/77aPpQwLF2k/LhXm8NY+HOZUatv+R8nr+Chp55AtTRslXvet0KeuRR90VkihL1Rxpy47ei/p
SsMGhkRx/c8WupVknSWr40cPyMvQpP3fpFyy2Xqsm3NT2IIts+5Fs61yFYHmGzjLx6q5DpS5yjJLmHc9DoXLP5DQY2Ctm+axeolo
xzGRNP0aAwMbBDGTb6414mUhm+i82t8Y3ffUkwZcPqsgs96fFcpmYzNbeQ91EyvFH9eTC96WOCM1L75Xe2pyI4mH4agO3qAZ7GCY
GvUEhqay4FnqIud1QxybVWCiMf/peLm50ssgaxGL5FU0r1e/GFtF1SceMZyU76qr1Zx+0NUFVle/uumZT5YokRODA/g5zFq/MHYT
PRLEL06k6mW3yU5sN2OcJH8irNlWJdI1qIQt5Ep77hgbOP0/XelaeGd2UOzHn0ndg6UnrhdYMkMXQ1FizaKTrnHnGHI2D8AuH0Hs
/Kx4GhZjF/gQRWtNjXpi66smmvCJzECnYm6Q+g6/RMDO/T+GLJZfTAfBtW+kI4oOqUtTiWBFcT8bNwcyHS5gzS4peE3xKV/5JB4d
O6oybMhEDINGlbj3keDUwgs5uA64tSHAAkQPe0eDLVY9YnayFJsaYGVjsXXHox3hsW7bAmnVHyemNDIdnxxzAG6/gUOv12kmx8VU
REl5YPYdZcqunWgVGx11sNrPBkctLJKJ8iGEI554c/1A4b+gk/5yONMfrS/u4xGGX8toNOP+vLfgyUwlWIVhql6RmVjaWvYaf3Hd
i5zziRmOYc2V/C9JpwpPOJZAIAL70sbFDR/tIrojnOCK2G0xUAxC9T9sVPpQy5U+aCMH8I8Btwi+HSam4LdDwP5s4NwMFYwOnTWD
Doks3LSs1Il5EvLrPP+arixSyBm198PhMmwdwCuy+tnfVaA0DjfQTQmkyIsLO/carKXWqcA0mCPInEQAdYvchBt4W45ifCVMqHGG
Nfq0e7rwHXr/bhSGcq88V+ncpxWtWBaHsaY7Hs/0cvx1WsBSwlIp09CkPCFSbRt1qk3HKUXFDnuv/16bWHNgC2LCwzFxhQpcluzs
YhxL0URC6LfZzqF11dQ9BQqG6rHXVkXITFxYG5h+JMRbQRNms8T0bFvVGxk99YBW2qirh9RDVOK3P0bn04QeQIar5qy2kKX4eG1H
Mp27soo/e+J7R9MZcNzeGBPMHrYbx/trPTdFp1RSZTaTBhosHiKQUlyfnv9q2dnKhyT7vq6SA4f/qgGdXkBdF4jNfN6PY35MOLSI
tiSrFTbBaj1R8Z2Rt5/eiG06uWdJfQWuqTCh7OnZveiWJjxNCvjbOpSGwC6P+2dPNqgWnoGCfn94Y5UX/+dUd5oC9i77OnRZisJB
o2gUpaJosOWYDYwWucvc7wd/YZHTbdkZFNL9KbdPkIXlRCuprnC2VoriDp2GNX7Qx2M8pOewpTkvJlJ4yxpkyyv3m5BrYEiCi2Pv
O9daUldvkExKA7VClDQQQsod97TLCniARilOJkKSkpJs6mD8D2o/hpbd1JJrmvZ5wFxDaIT/Bl8UyLgHKOHpzn2ewc9bZUZBbuqq
9K+Yl/pZiyA97DeBVn8cdOCgroobedIhL/yn1V31YSg+8Z+j/5hvjKgfdiKJOx807zh3n7F8ELAJD2/FzVpnDLVWlWizv6uwKBqV
oTvgAUE086/R/sD3XGOhmz7X5TOr2DgzrmfM6ehERECFSefz5L3j1m+KC0JYAshOHgVPv/C/tEUXjeZXriWsSooe2BF9Sygg9rlJ
+7w9TpHEmWVa2SxeDSvEf8ishKmt8pg7SA5RIIFi7paxXB/9BSurDen24J4VzPY4spwG33WM3gUmQeT0Oi5qVI/qJvJdpDd4On+a
WL4r17PBF6oBv6mljS1+KJCluk9gPqAt48PGoAm7mnN+wk1iK6IFLjnZwzlEYfW0tKFmbHEFVvAe1hjSVjW9xN3Lt+SI69a6oMpY
MVXnpGab6THhhULW67rdNiSsRG+0NacfVP67DCBoQBmJaiASp8u6QI/f9av/Tjj9ZFOEEfQoLII0d5d1RJcuSeIKFsxVHnKhvrf3
Vi4RoQ71tq31d6c1HQl8yzfOujdrVNztnqDsFuuuGlKwqBjkEGxBH0Qj36z0s7xONVeMKdGnlqmyJ0ad2j8dR2naaCMeNoQxk8sx
c+XDEUmbJvAdiUf7qoWtJmLAbLBU22bXgb6vbAXDcxM9o0Z7Takrw98xq4nVGYHPvVKommy6b8Wp1X91JRB9tl+seyliKe1CeE7g
rXHEImRfuGbJ9SA88ZfFTaJM73Uz9ShMSoYE0BvX4FQj1/J2LkxcB9de+3tB3OsVudYniA9YZh9v4ob1kIP6jW3Hg37rgPqlnALf
Io/vE8MvA0Jg3/fp4x9okmdCP1nvQ6D5kjgKYvuT23SqJ2hFXhYl0ynwxfXSk30UwdOmFdmLn0vs44NVfq4oijI5XTYFMWxqzBbB
wryV00hIjvkGYDphm1z2aHF5TXq6+qTVKyS7vBkLSrJD+P42gXR4K+/8HvbRzc8+HmbzfTx79OqxKvOzI1LJklYBuIgSnpdS/32X
588AVpXAN5LQC6Z95FfA6TBye22vgCml15hTtSbM7QC2JiZT0y/LjQPvxHY6JM+votwlOFyS6ovKM6Skr83idTK04Zuh7Es7AvZY
t4J1M1PnZxr842kcn2GXCX4wIBrzPB+7mIdV7E8FK/uYLHBbIJHiPrMH2q6ka250rVnnctHG4/U3ezgli7HdX2Df3CKbHJ0LOtcM
XpIVkjaJJwfH70tuK5Kc6/tkoQmecqvGAKgNHqiTuwgic/iSMtjVDLQo72g3iULr4UzoTFLjxfZ2L6UqQ2uwau3yw0xC0OHRNpFc
enz0Q5RjAHHdpAyTSuciFcOZexNOc2D1fpX3Lo2j+wkZAMonREHBQxpmHoGZ+6R31imEir42JQ/nSbCuDrUs8dgp4490PE3WaaQn
SC/Lab3rXjVo6xL3xY5mMAOArO8nS97Ho09PHE5XzU+e5aO04wgu29fpZG4mCOxiTFp3T7GtXPs/95WNYE5FBCWFESryv13kh2Rj
oHjqW9McgdOYRu6XWOHSJBaGQDcIdSbX8FXyw/wAiUfzK+FjMEmxQbl4BB392nI7XKBe2WFlYtNX6wjOvyLjzuupJADv4ntuKhPd
6toLhM5BKyjzwWndpyweUlPLvv6/zKq7glBidtNHbxemBk6jd7ig97Gk8umRRu+baba/2CEA2cjhqx3txXY9iIV6DzUrqvAAUFb4
+iLSiettBQfoJpON01tkDMqIbnPO8KrY1B1n0uiI35TbSIDup7Ye8J3anjXsVi8S7dsRDod2DkRErGXsZAsxSMfLbLYqFH77KkDP
tHHzOiFnVOR6hdGGKWCE9N3AXdlfWcMlev2jlcG+Gdhjy4ga4qMu64bLkD3pTRrYd8VwUIvGuhDzIcvwDjRT4abq1AtwmjHBMimB
ogSYs5/z/zmqDkjyrTI6b16duD7jU3Ojacgmx4NU2uvVzjaFHlfgSl5HcvoZztao4qQcfx3HkwPN/AUtzpvdKhkD1UN+zZW4Ug7h
wkS84L1hydv+arCICz/KIEAjlEyyR7sgGFXxVHhRwoF0ReJFfQjzX17Xgv1NU5XNIy8N6ddOIiRRSaRc0t3FRV7p6bUUyCbrpfuq
CJhxjFfwXD5yz0KkKXKje8FngtCRhbSMHM00LzxCRTWB9lOb3V98t64CMcZyAKyxdHPOOtXbumdTJ0MBvyhYQsA8/mJ/LaRbzyoF
R+OwVK8mcm/WFvQDjab22W/Jdp5cWK4Dq+VZMWLADCbWwie5qTuHnvbUIxIOK/MjwIT3MVwqem42Qjnw3Q6+xIWXvAgTvHmbggxm
KPqVx+2ApMKc4u2qwHpHo0bezzbNafY3J0+dAv1qzC5OLlZG/dXBz0qw93rbQQgF3Ocm19sNSYDVmu8O2iGI4ylGPogX3NHYo2JJ
DHVXeDf4icFJllZhXWkCNdarU3yye8EmkMgzJHUAjJE/2X5Jlo0D0SxdrzaquEsl354aWquk1fuahht2dS+aMaC/z/P7pmkZ3j/W
XkNGWO71yAmckAgKO891165g6vqmxzfWNCbAzvaLHRuB9KbCjroqef8FzsEPJ2eiG0NF3St3HriW8kvAOt3BWZ4//D3PvEJCasR6
o+KH7x7o4Bkic+wZZYjd2feEHBgygP/d0vVcYnqrsr76B8bSrNxZzL9xz+KHVRhNCpS4WAxA0hT8b0vThF0dcFWaDc42ENBcBip5
6nLmWvP//CXLoACZvmh8sTQUNklAQaSYAEpAUDI6BQfnwtDcP0r32aeQsbnuR6i4HQJaLOsEpggPiaPu9grZDOLE0sIMcOQElPsD
UoA7nLvLTzORrtbOLTRztaR2qTnYFJUeKXTzMDNzfRdnJFUvnhxFfZqrRwuisCBBDf2Z2IKhPo8SuxPFH7CaknuRSrsxw923tKIV
XnQckYtoNIkn9MCBAD7Zvou7UsrDWF69hkyit+Q533xb7/pqS7KAneu6iI8z+tINjw9hOn9KIjj8tCtjksIW5YF1/f1FBcKClxdZ
uTRASZW/evaaKYkWGrawgO3lWewMh/TmFoftppdvj0OqwtbOBykl/+DLVgB7g5DrbfW9Lz0HmJsCPeZzM1YSRGLqtgSf327aoRRp
UbgoRsaow4qH75HDL5pguOdmLT/I3l/Ho07/jNfYbdtlXHWCaedaiWFfxjtQHbMSPTouCy7kPUR2CWoNHSzl0aRHaOG10/6QsiO6
cg/wMS/dnUkZOVCnZEIGV9jNaDoMJgoMOojnL/bIy04QBmMcoJoNUebj7e8buicgzqkidYYGCkh7630hVch8ytETrqRquO2wBLyJ
lGgJOjKx5a/6PZ4xFqzjCHhsbql041Dah100cmhfDkaXh2swhBNSfcgTczKoxKySfln9L3q2qZZ90uO3x/xPj0xNC3kEHbkWc3n/
Aa4OTrG+ijlKH5G8+rV8kcBlW0fmuHXwiwTg9Luv25V2erv3pjHXcyGhRiPHOsNQylEK653OK0q4SXpoplkWb6Bsbx7EAa+0uh9Z
J2qh7f++IDpHjivU2PvwVKnnUJmErDn+1VpVxMjkUTFGNCP7fYKZ+yzgkNifI8ZL+ODhV9Kpg5DmKRglaAitjtilUWmM9GDCzG1v
a+lUlmoT85Y+W8c8FwfywKKCVwtDoVQJFR7vILfbiat6P2C3wKaOpxH1SgOhC9na/75RtPS7dBhE8V9argbVsw5gEoMuUSH8XY4e
d93gjJXlfXSp4jP6C5pbtQjFvEDzYO+9NMr0EBoyveMg5uUlK+a7+6bEaGun25pwLlej5fzKzkiUFZ1XfxttWhEyNMy8oHaajWw1
HhAuK48+NuypKNlQuKxv2L4IGVt9o3m+xO6Z9Jjx9alFXhMEZWd+GA5Lke4/EFxEQ43pe9tIwGrry2ZmLb3w7PWWQ1bW4vo5JhfT
QY78F33DNGS21cVIUb4KLQLYOS1X8m/cdACsRnZ3dZwJffuWss+fGUIyUtKoTtPI971hiweTLWjRFXnwYvEc8MdgAnLvrmvym5tV
Mh2KVNS9yw3l1LUkuHWrrDJBwfjURxTC+BrviTpiVuTmj8E8n9aCPDw7IO5oYSWfIMZW09ZFK+W3cs2v+87/pn+F0+PxNsoygeOG
ebQNoOTNGIGtVXfyGXxlRpGZ7xzgR8mhY6e0oYZC6T8bKwA8b2D2uzwpRpfupfRGiAtB3dk5ePy8kW8JEM7jUw9kHoj4zftamGuU
4A+sEr7xDVMw441LigIlS4Axezdrd68ix+J5mtTtTXTVuMEl78adbLhT0aHxa7pph5xtYJhHCR1/DS+3CCSQ4CuKJvwRJ/EnXXK0
UJeCRAK3Y9uzML4AWel5Y9zTz3k8OcgO4EktXsJZFJVBU9iDNM8lcJiAxv0ocem049rVDiwSXYlgGgS1CrUPmS2ELvHIVBTSMjXT
3hqFYPDnt0jtoj+pU/5meDFpwgadM0KLFKl7nCJrcNKecMe+pDDWAFIKd5fwabCTuP3LTsabCeGPJ1QLW69ULlqTawYGWYwtJCka
u5BXKo1UZvn766D0JuXQgaCcuIpKRGeM7KLk9od+uJZd/4sZofJN1fJkd9zpH4vX1eD/v732hPS+d58GRzbfxdy9qaJfGX98RYvh
pTU9mCDGC0ZmETcjqSTwJsM1qLixaWYLFPSdGie29SJyccdVBvLNSFA+4kfLlXuwbv36ugvHpiM9eY56Q+ppMfvxRnwlmeUj/rSe
r0FPlbvOGtkMYHxgiaOzSHpuA6egyB6wQQpsDZD/pkNB0cbRxf9DitqSVboqib6TJ7tfCXNl/6nXAwtkTTkd92pscUgWy9VnyfNT
+ZnRYzjMFzGGXioCRtnFv/3CJyle2l60Tl8bOJ0hoAWSl1q9KF+YD8RA5VJnBYXe845du8TC6nUrAfOPCpimQpgdlFpXXJV7HPJK
fC37d1BtorRbZt4mckM3vJSJxM1HG+zt2DoQUtOwy24IV0BXcjLi3Yyvuvg3h79SA2hG7Ux1+26R/tXaMgwDem7XhKiBiVy5mjli
1zXVhH223N8bGKj9eESpFyRbUZS9lPHUpv2xi+yImXFHQ2rnhPa0wTX5lcLeI4enu3KbJE7WRK9h6eu0nu3yEXQFeItcwyQCH3mU
QzpSG00JTXMO/b4bGEz/cFKTJlH/KiZVupufqnrBO5a1WyHhavmPOnWPqeryGoS9Iyx9qNUqHqOfuJjxrmpfhuLtgCqARL3L+7Km
xosvptdfacUkuFmCTJcMYlmWpQGVeIst/PRPWPpOJpSw8W2phJv1dovQmIDibEnn3MJc7nYv8IUigFADmuEK2e4i0ODeD0fIhXJh
kutWsCu7o20/ykXlIutBCCrCn7TdKQTFzrZGW7A6K6ozCQrl3Yv3PoLYi8KVbw+pQc1JBRz+fdsSUYZ3z1aDsgsHINhihnmMio/C
kTp4IPaYX0vRAZa6Dsv75vHFX+yIdVtd8fHmfiSc+0q+q2L0L0UVh5thCWwKcJrcoQr/wa71sLVCUK+isvsWohrVDr65jGazGeFR
FjkCebpERxVxxl19oIzt/sPp0Zi7Q+vvSzdkaxipd/tb7XEmCcICplPCKSny27HvVUMjz6aJ551jwft/HKQsbagHXJNh5dqoquXV
MlGjd9dsxzNmu70VtLDf4Lq9S1LIBhqJkk+6uCyAnkzlduJRqAxzETP7ma9APsawbqwRJNByd7YcpAl5dH5BjvzKy8YHQsqGrG08
WMrgJE9SmAFXVan5eZj1eWM2A6I4S9AFehHhdQDpiV0TnDgp8/kzcFoSVOJqFthdaZqsGdd39iQJZligQCn+Xkl3IhaB6AzRRB9k
E4mt6ycEm90G5DjhDT2opacSw9gEs1IEkQCAQhqTAuTZCURGDa61Rp/SryvfkcuScknDMMFe0Xewv59B5TqLLWSCJ6XaERuojDy2
HYjBlcYyzybDDz0Pu3XtSOMyrhuQrnqjZZipYUHGkDGbMeFDzVZPlFFjGTkLomHKDqQaH/ov2zunDbqnJGv9CCSHM72ejPHkpAWo
uSzNcZGp0lOhzYve+FuwDTTUHY8toNKzIp2KiHwfkoHFxMPTZplG+L0y1/bucfIaILZcs5LxSGsLzXkChaOzT+8PxeFO421jte5/
2UGhqBK+9K2kiA+KMaYkUC8cFNzbrKOqxrtcoy2e5hvrN89ktYhUU7RIPdhMLDEAybESpeSzMTYCCXCslfycLXZdWtqYjPvxehsM
TKVvSEYPITCQoKgFuGIjQBE1U7zzqr3o2qKHN+m+sHKybtRhJZ5TJh/JPYLfUkG7eWUZXmsIfaxHv3A2aDrv4Vcyfu/eTSU4HLPi
Y20wa4zSnSNfrPM49rohmM5hso17joFz+jhErYQF+mJAhvGML8YKOHrF9ITelcERm1Ztka5HN1JCcXnU/rl41HGFkTDwfZcUzsL1
yA6kXw6imvp0N3hJqeF7+tWx03v4wyIIU7WhI5zqHH/PPPMQ1zaM5psYJ8u1Ezuf9rj61iAPHWimxufxtuHusETvG5f0w5oHr9rF
Q7PTDLU6VKeXBic6YBzh3onOnJzCSfsTLhdhe2Vp/PQlK7YDpulDMnGO9IlyAdJWhbMtm1Fagvo+o9RhfVgtDgFNDrJ6uUXPZ/mR
SpeN4lJk1S9uHAwrmUNYHhgOrySE45ej5CkpH7npYlrA4NVLrpl0+MlcERQZ1csL/EuYPV36WQKyMzoYjx7A+5c45ERkFLwNsFFm
yVKLhcpShv6RP7mrsKxPzUnb6qkbLvw1Jg8pAYv1S7FMP1pnuHfCzELw3WGc4OsvvFaSnCCHtbIm07JlANKXLPevAqP5naUznyJT
Y3MsGAkS71cSNYljVNXUYdMVlmXa/4ITx6/tvZBNlvAyUNWH3fYkR+dJvjn76f0SDnkuChzlXiI8seR95YXQsIIvMlETHK29VmsQ
4UdAJbdFZQ2YacnLmqvlZWvEwkQoKtBtqxuCBWpUMOJSkdCIiP9AijEDVvQUE7jBVlL2fZ8R5qsDvPVFULI6G+GbipWqQW71qk1A
HqEynqubpXYKfTNS8bZfnmK//6pgTIU9tOplMVxWNjVEXHWGH82NbDMhAiacDqPPPd34RcLalEu3S8wTqopJ/UR8mXXXT1K2oyOF
knBVnObkhQOoZZIBKWqu3Or+Wi7sVOmAOhHi1u05HE101soOmbL8SxHSddaDBIwifCtK2bSF86U0zQibuHAbORJ5c0KhgalNoDIf
s2fKD924yLx/rYU+O4SeN3StPNiZWRZanmZny0g5EorQvu3NmiJvgODxsZ5ldCHdciPZZgn8tzqBx27DXDU4QiiC0x9I9a0i/+u+
H/K1gsHvwyH/aHUZlXu2a2d65VCf7lgwDb5kAjhuwB3+orMjW6W/3PVa5c0X8R/VsuY1kINkOSsxEkrrh8mufHBL3kshld390WCQ
nbJF6h8oSM0LuGBYFofSWua88rcp9TgceaQxdyrMTBaCB+S6uq8ganHFF2vBBexvPAt1noi0k+ZKYGQkpJ0BMUknra0i9pSlrMLr
08ZFa+/iX0d5v9k2tE3mAYDdH7Ir9Z4Q2uI/y7p0n5VT6kt6qKrb1vTqP4eQm6jJ3wyqFQ5+FaZRXuFgRoQBUE5c2hFR6tHh/USU
uLzrJTdJZGTsUz/86S6/tuR9sEtcM0X4IDrHJkRRpXYen5YMzeu9v5BAQeyyB4yxvvMNVJF6F+Bbtvq+cRod9MA+ZnFHgYU5MPpI
vVORJlRBW41GPkZI2qr9m27aBCgROXqM7GQhPRpKhJELnqxtEwtxyOyi3bwQ/ivXJPXgC2t4f/YDHbKstKkFNdR2Jx1VpAywR+PH
cCnigrbwfqTtZynPBvdhxcKmLJULgc3GYM8omhdap+rgM3ucCQ+NzqUcM42S4C0ij56078i9v/KAyJ/EzV3g61K5TVNVqb/YLaLb
0t0sOKIsiv0X1SgMIPXjrw1xtSViiMavhoNgywIDAbPzx9xpWIH9eMyfBv1eADK3fGJcpJ4SBBo5hhSts/z6kqy8pkF/+eQJe23W
razNZJikmL5Gk1/w7lqdN20Wfu0LIZl8yKhHnpuu8Uq/EqSqICqM8AcL5TJpwzugoXE356duwd+E2psbD/8Q9BIBi0FeTKh2e3LA
wKNmbRMi6rXgmdKK6rzqx/xlX4IlL+5L+GnxuD3xcgs/XCUWxNPw0bJi4oX1b7Tl2is27d6euAWCvPTFJgY0fbZN58wAd/1W5t/W
1tdCqSQuIPtC1NP3Z5mwbZq3BvMv6ssllysttMhAn4bDit+TueR+iKSNHlafXp9qrJMKpjhqMkUjUp1gBbSxCyg4ZQKDuelTGpsC
wB5iDy9gePFgb9rwWyh9oXDw0YlGal3t/0un9igLcZSeZ6B80Ot4bgJoF50QhkseyyTpsbE5FxPZnMv5QbQyh7rRnGzwiVNlKmcd
IuP42Lbf16kap7F5IAzoJ9tLihJIWsDn95Ty29IgJ55D2Lv+HIFQXJTDn1MzR/0nNgqN2LGgRLBouO54CjBG0fgDWk7fVh0vOd9/
83iYgWkAFSatLwdwXhl/Kv3gS/SGtTANq5cySjSKOtx6S2MOW66Wy/TUSRWtT0zVxtOyhJTjCyd/I2JRX5KNkL5WB34/SpgEsnIZ
3HTYj25jCVIKHTvqo3v0BllNA6CXyuBJFqKzMe/klJn3Pkdd5w5Ry2wxb1XkpntTluoH8F33Dlcll3lcoG08ysvTBng9FfO6ax71
CzXEOQfuK42fNglSEtKgIdnbDRixYDZigKHOOsfkoYpaBGE5pLIey88awpGUUHiz/7syma6RLgssVq+8SMgLHS6r9fMWk9gO4/H0
+AMyrDsZ6b8uyd87J/KX2T0xWB8iIJ+6kM5UwT0Ks5CqUyBy24Kj2SQLs0o5EG+mCqbscsB2GpmLPmPkqfbSfVndbgGTJk23uMv6
zBTxbSep+OvF1+A+A0Yr1Q5Yi5E+HNCg4GCr/zQ7liLDjL8c5XfQaIqdrUdGqJFK4TGeHpf7CFm/uph9e3UkTRcExmNropGdV3xs
hYdgfdwiCJx9xvXBijVukmg9nLOg9NfBgs9U0txL3GAOf7tT1LC/GgTG/du5NALX/Hc/2H3MkOvjgP9suBp3HbLBxNv/jrmlncYK
zct1JhWrKl3cfrtm2eQJOfKXO8Otywg1noNm9SYHriurHzbf0riKWQK4EEYvKGBtxsg7kZDPyw8yDGLibQZaLckDK0RdwWX2//B7
m2WOobklbhMltgy9VQx7vuqqyrzbqfdzMiEVEQcwo76HCbnUqngoE9WhzrC7SG0WxWi+LIlRrjJhp+9yD5/qAYRtBC6HLTEz6KGe
84m6sCzfCou0BBYXDulCT+GEQIOua5XOjeSJO9Y9cg8DwTwfqYbP+W4Jv/KcdkDtiCw1Uc7fPk5kmwppNZ1nVe4EiUhE5/nJFukY
zpyJEm5G2mOj2e8O1wkexBD3CtAveT6MefMRoLQlCSzPeK+xc3buvlkMOWcZKOEmcmaVIWmoxqOtjbFzuVzH/5NQKbkU3Un2ORAa
5WEO+YYBdXmACahpAiGeqXHOlmbit69Uj5o84xCzrhKdeqCoziytowpdj4j+faBoQMDRVb4AmsvWbdm03clrQn/0hGHElL82QN/O
yaAWvAsQANM7Aa/xEopWkTV/84jl8EFBPUmTEFqwD0+wJO5kYVZ+vY39YeAr/hQRPpTaBnJrXt5s6tsaeLdFWj9ftuu7mtNos/5p
qYXWWz+J8dc02Oa+b6z7XWex62BvuJw84A/Rf3I31S2pFVWhIuSQ4/Vv5bneT1do242xoLDErqscERXqCf2D4fL1O/3ltc7Fn0S5
vWBfKukT4Z5SyWBI1NI0GEc9gc1xzGVqgnz+BEJM9Mr679lhxudQZXQX95jr+JYTREPXo+5X6WHbT0oQ3T+Bqn82iPDSU+cwQQDr
eKTuDfq+GpX2Uc82UYH3AaAGeceuda6GwS8RRuPhApPQsokbMQAybidH3jPJbdUmcUQlupPQ8KlQgqylfl6WaCowjWKruoJSYWII
fX7X5TcFdHvm/Xdzn73k1k3rwO8Sw0u8ZHf7PRopXL46bTQl/+A8sghaUW/SROHfP18euZWWra4z/OijKuwqZp5MK0ndl9GNO7G6
fsDNdvuryswddfERks/SrFUy60Sy/SSjdkV/b2DIigX+OrwbMbhsDk6gm3t9cLh/My81N/d4apzf+UCZFnt92hQ9CNwgUCB/pNFr
BJKhyJRr7zRBPeeS3NdEEaDgQEjte+2nLaZpW3Vp9w0VXiVyqAm1BgqFR5cBft4FRzBLRBPiNUUVdWTthbRvNr+W0Mboc0aLdtXL
26j3gEEHKGFqE6KLzTIYV4JcolHQrEW4N5Vcouk9p4uqj5uDQWfnRLdJVS2BTX94h6hIWoqWxSnC5Ze1hb5w3gc5qFp8xfuW8nhW
4oiX5nd/wu3m8DAITiSByB8CpYMIHBqKBCdr9MH3Ue4ay83G60bSU4lwUDhBgq4Ed/uKECrrE2+XwhT7f4XMP9X9e9gFCcBq/IuR
tuNsgrtbJvxK5bWxGQe0VRzPymoGlCSYc4HxI84Q0J6YhLwsvkoqjzOuLxn61kucL/HTxDp9tPVxNcEDQtPwk5XoEycmUEHGwa8a
9rxSSM0YB3n0jSbplfx+k9vMORjqLygYqxJF2ooDlFjLLo5i0Gg+BlpTxcgqnu6PUDfHcTAVLAi2RqwSenw5Yh3joxwwl2RxFbzC
7YkBeIvl41h+yNEcWleO4l1tlzxvF0SFsdlDPvbB10Iodg1D7AjP1dvQVusm5fb6GQDxSHwu2Re6NMzjVFPv8lxOsmyxVb/d1ng/
23A/fMwU8jVXVOzIaLZVSr1gMfsdGVrawblVOPLImCyrlyBdPTRitvFeM34SqiGW69Wp5sLEh5oKOf03nKNC43B+n9kmFafO+8UQ
SVa1klY/Kwu+T34E4vcKotkazq5xj8i73hboUCglDzHcZL85TE54GitU6ZemZ+QKMdIeutgL8ukW+qBLx1GgJOsl4S8i6yMu1ZT/
cgECN2GUWfwylnxISJpZLzHK2V4EbK3bfP4dyzsGPoRNyQrY+TppYcdsGmedsk2Ipj3EofOVx9ef1LGj5qcusXEi2e2NI3DylgsQ
vYyqnJBGnbCbFA5wjxXKSROztccSFzfa49cXaMEcYkE5WMYqdLpfqNoxTQ9AXZenog06v3OJSklxJ5SullX9Ytmsnn4bKIw8yCas
pgz3uHYWWNK6jAI9Ls0UymbjrfsqLEo7/1YbzY0lLGjNyrsunNtbemMIXvAQgW9y/QiIWYT4EfmJJQURO/hj8wXgSGEte/NQTjdM
SJCFQMVyVs93tmu1yQhy1MRFaq7yc3OfEbOnfTP3N7mSpnHe46PlQUS5pqlCBrL416ENM1fDykl7KXYcYtAgB33Wv7vEZmOs46fN
g8X+cWyx7RxZ2klsHYG9BQjYY7/0wLs0iHqZEfOwEDL9GgxNTqDEOkcQaIbH703szMwIH9CAzDsVfSCzDgFVFNXSj5/ox2i2E8xV
f/JSQWKll7o9wAItbpqs5HR65Buy1AgRIQf5+TsUn0vngNUGu4M4TV9g7gAFh2XTUBeWfl5OOFsoftVo+dKPSUzWZBgLbXpAIVB8
zJ1yYQcfEen2sXxS+r2zjAxOyLt4bIGNtis/H7b1FkQur1Xjx0MyfeaoMRoDx7rOpjrE0AZWg1stADOrebmsWAiWVAz32Q8Hx6YG
AFJcAZ+pre+j5PTjh9Vc5vtW7aFP2OWtxj9ie+g9uFlFW1nenfH9QWoMyXLIuitVYITuQ+gSuEHGQW9/sDbhiq4N6ZKAucdc5RSZ
xK0Vyw6GiIJr7/Xwx8m67hT89XrdDqrnzDDVt8GD5sZJziyMrOyNHlr1UHAeOBQfpXT0GLbeQWgfD1By/3FS2z5cNScK+OS37pDq
fdFJOhBMl1vNKqc1tNvmjaowd3UQ2TGmu3putFBru0LOTa+K94/a9Aw1s/uxvXG9BWxfQxdMaTtD/okJkXdQCGFelSWVtLR6Xk6w
v5/f6h4IfBi5L+8ESsvLq9ACfRcYhJRJiWXUhpAgrBa+wMPliXP9tg8h9osPRLnbT2XuSC2yh1HDK/zzfGD0TyxfEC54qagbfvni
JS07Xdo65p5r3KuyJQ0JaObhY3BOnlLh0lnA7G9qDI3XTPHwvZyinY0Ckmk6kQBLQaGKjSj5RnpAmfat/vMX3Ipw0XL/7nUjGk8o
8yYct5oIA7I32vwzPhXunr0RKsiZELZ/nCiVs0Nr79xkOexNhy4RcLwZiyi44I8BTjDTKUClIBwWI5Ri130AGDHCxBghp38NpemN
0uSm2p6Sc2YLyaYvB1RlW48/a7fd4rqxwzHu6q464WOh9MSXFCGiwFM8ZfVCuHSZZauxSfzy6k3hroWX+mkZ+v5seOd+v4R8xBC+
4u3ZVuBz07kI9Sh7W4rNKEDxKUzsldyJ3trexHMNLs+mMnnNEtBcsRJUDgTRCQKX0UHc19hhspuEI3gpaeoCmGUuXB7eFGpe6xAx
VltZRwTrET/pldONkvzhWuRuVUndxuC9zvEotED2DzA2SKLutjg1ZgnmUnfzPT8pujuM5qSYxo+CnaV1ZAQk3ejI1gVjqzRafxKg
/9T4dklOwmUyih+mgreqn/zwmNowgnTOItiHi9ExdANIQ8ROOyHZB7UZKN5VQCZOwV+nXwoQS3m/Jk0lsFqGn+IvCDrWPBobA2iJ
d4+bjxs2EkiHwwf5gEXlGDJOeBAjXVBo1w9O+V6YZ+F9gb4E30naVI3NS+G+TS1Oycx7HzjVmtdzWBqfrDBBW6vTjzY9vOkEISav
Q0QXasjcud8xxwYIyMbBCQuv+57C0beW5W51R6Zo831IKth6e7iksguyjgscFUE3Jj6K9izoPAfMuwI4ZnP7t/UVunDLofCN3jSa
PNDvfdC2aGCpmTBcPlp3pnzrfvJIBFMzAqFYiY1tJCgAD/0ZvGVPUEAbquCPAicedJSsZHmbAMfZws8r2kqaZXyj6agT+4SuY8Gu
oNeHMe6wEBeJWYjQ9390cm+UIn2BPEUEwUpU0a4uIv1kskDbtTNMpy0lV/vMS5iZeoBEkl4pghHBe7Q5kjq0AlNXKgUAlkXuWb6S
n14YgSRbVGng3eGjiFNicmZEt0sRPmxBwKyed5DgfRC0HCrfHb+0ePol36g05X6mLcgz0hqj3OVHKC37xmOEyWGCBK2juK8c2jL8
VYDieBpNSasO/3VYISnNYfyci1b4MjfsMGgErGSD6ws2HPByWOW7QDyLpys5cWhyFtIuxBYNnvnPw5TDsF9EtxyCWcwr4avf84Xu
nqaEd1HiX0FtIAaW6cf/kWEahXKJynHZKFJLZgy7c231MKqO0mzCzwiOozFETh0jEJuvUlheAf31K+vsVPYXW5zP/ITmyTnJ7VHs
WPdGcjKO6P5sHZwDDVYgYLP5HRGOlmceEi8bFTk2n0zLTwD5gEVnKmkf0PZyl+btnOIfIA6RSR0UC3KBkf0IUq6Go35NVbYHd2Gy
3S7g50UeGRLVzfnfynFgyxgcjmprgOA/g+i4NqpJILISNuUlk7AC4zukKjDagEaxbo3HhUE9wSwWJ4PkYto2NNje+PNgqkEye+Yo
rx5GBsQ9jiO4Cl2FfWJ4xIPtsosERg/IEzsQXY45avt3PypRT0Fnbp2voo8jgrHl45qkBEOt3Lyld4pSqM5V41yWn+Xefm9VmpDF
6scM/77/hibQei6C+eYNrmGbW+iTsb+KI67JTUuDSnRCiNSL92uR5P5/crLVc3T6TkkRSQPbzTOWwtY4T4/t0uTuQRovaYvWRyUK
drm4lDCW6CKUFerRWeB1LZVaYZejeN/X6K+Ss0GGVsG/hwAzkMbzrCPF9ZjtASteqt0fUK7ATjtXbRWOTFpKf3dxXoTPoZr0YrjA
4YjVqetSdXKiS6Fj12KhzJS3Tmnf/mPImmDG4LTTVg5stHEPhz5dW8edgXURv86QOicMwJFgY3D9xzxeyuMXX5JfHQTzFq/ihu8y
FfZjQ32uasLlYOTz7ITmzTjSahUeHwYqQhmS5k/nMsigQCvR4O+7GDLUFJKgxkAZwinJc7jkpKW0idRZTd6LhweFg1LyMp53iibe
pQ/e4S/bS7Y36zX5Z0pRnn8loGHDaAIhH+oKwkRfr66y4K8S65YVDwxL1GRWghW8udNFqeQsBYgxGKQus363ORc4u6xn1CDESr4d
+4o/VBDfd+38C664Y0JMsuFWZm0hJulAFGZ0687FYFetUNnAonAhE6b/djFfdKZLhI56z8+b/Fzld0muWEfonIQlkmksCGVaPrKt
8HHhUE8qHwn4F60jkaB4N3pYQUlXAoIVhJJGqcfKsQaPXBz8xyyzQEFZZYacjvYm4bFT/rilerDQhZ0LNZhi8kT+Gcvc+ZMzCOtq
ILHV6kA6XfXHOjMeBR2f/rAEjRcPOuKd4ixPHnaenV8m3M/ck2KRwy0hzMSP/CL2pBt6u/awEzFOsm2lsg4dD3ujFFt9GsnneiwH
nN0k77awWo6WTYgeDRctpCNgYcmsQFR5tN/+sgB/5Uh067OFryInk0zkm5piUP59godmMIskr/LoIt2fEq3Eh3qlPMpW8CUYL2Xv
ZwV0jcXsbO85adCVeLVmjNPGRig4YrGO2+PictNNs6fkd5mMP4wKegWXm3dC+Y84ASaVWnM8/VU2uog0PqCX6Plzc1YGxrsYdiut
2bdp0jDzjpwhfzCbCKGh+J/9sVgnIwVV+ctcZys3i+xWAbrVTjtFn/VkkYA9rg14Yj4woT20rvCCXzlpAGuZmod6CYpSmI4Wow4H
/DUo6QM2HnBLBNPoRO95/Y9+0f1F3i0hhO3/5DWhKbGtzz/CXZqo75ivrhz/D6FQW8ToqOeL3dzoQnRs9nD4u+mnTLLZfYt6HPC4
vMmY+n8luWsg5sBnxkbs9v1i19bsmHqve1mhKBoJKyAY7T76rVTRb44J+me3Tb0+uvmkabSNTebApc6KlGZFJTTvB212X7fnSFuO
VRZeABILpvELuPNVnlitkD1gkiXkx0ZnLQj3q0UQp0urmbvCPPTMm5miY+5WKuxmY2Ap/uKprm77EQdoWCikWPUotE1oknbpOixo
4K4tiVHlXhrq2r2ljiZam0apuas6R7VmAm9ce6scVdOW4HAV2w/4vY7iegbw+OOC4BvDMT9fqlmXH5qFg3Ez6cy323LDb+7qMCyP
ls7FkCjBUIMaO3Gzj6yz4PM24JCJZVntpU82+4hxNuOnRgMsYa4pPdjBWdSY+OvSFegq63RR6ZAGhJJM4P3U7mxpSPfMjbkKnL62
fYC9l4tsLQWgaqVvkLTEqXxEt1UIz5l1z7NIJsD1mq5NimPjTv1BLC63DRUqWOCFGTva4OxsZyKjPRD5h7OHa+Fmn+7u/voRBLMM
gRT3JkiKj3qPP3byarT6IpSOtICNGZ6fpO8dIwNECnv2ReEg5dCoBDZC2lV/nHzUE/s9SqiLx070MzEsCpw21yOCQWJbNvLjTtvp
lITeTGMn/dWg3bfO0KyW+hine3KvZP8YKPbnSEmh0vDfuX7YM308MH+mzXRaTVtShZ2rP+eM/bEdxjZyIN025h9ByWo/se/Vsjia
wAToD3zq6qRrtWheA+Rqu+7cw2deK6KJz/uxu4SE0H2Rk8JuHLtKFWRWOdvpJdjwcI/D9XWHJ3a2OCLOv8gLyVFeI2o7pKRt9wUs
uVferRwRVm60BVQM3kydQLtynVLqxWNa1e3VOMsT08IkmiHEBW1RxXcGORKjc8+FXvEzpbTQzEjS36i43VoW6U0TGT/zF3XVvCwl
HK4bo43+7S65HyZKypL5+Hr9P5NIO7/5C49lNsIp0GgenvqA7phGdovWOY0Uftpe/LExNRRVOyACtGF/vGBa3Z9ewvwLf1/moGu6
isBPdHIvAy3ujV+uFc0pUEPOgPiYAX3S5OKEz1fevsDzEpLzFhQuKx/pkmD66lizXzMr9SqXYDltPMzPUDJhyCMhYku/8DY9pWi8
cqnlDvOUFVcfSbe8GeDoKyrAmB1o7UrILW8L7b7O2n5FeIvs4UmEt7IEGippl5sGIYCy4ishM/jMoL4WB0CBUnigGhX5gywg7stk
ZkZtppcbVrCkkEJZeqguUi2X7hSAzPKSVy8qHWqja+nmY7FeZWDYSZSYYwBM/21fuSkuXu6MEsv5Iu+qgxPq9tz6X4ivz19RqgOc
VjMwJehbiBVW2H1uooDH86BgWENGGP5mEdrRBpZqWeR2sOQ6MkLMejgdXfiKM5vpr5su8O+gnoWDiCXupw/InKRlarBG7qGDJKAG
HZLSUMbRfvqdWsg22sNIv0Bz4tGyGfuIzC4kUCsYNzp4iYJkvqv/4ErKE5mvfiEj9wV99vrpetnVcUTaojbkqtrmtOL/86rcPy3i
oMUUfRtiMM3GYSDVSl8AB4XfpUtO7Pb31W02Lr1mqPODN3ekitphPkmH6iFvz/wGDB9i2IOkDGU3lsTOh4urx8uqw64mjTCukaQa
eqITLpIKMACgyknmoWTb4g8PhYllZ71LVn+5T9qQuC+AsobImQiWaqe7O/0++HuNCrzpuDpiu9V2suA0yTEyuJvI40pnWcS/+cl/
FKTCTn6BNaGJ8VwGOatv7YIneQjosli2hfm3LmZ6eQfFDQnfnfbCj4sLsGRtFFeKt5Eg9B8X/HeZyewyOXAAudf2hq43JSEcU1R6
AVbTtNTcs7W/D+xlsxIV97s/k8FKqoFfX68ZZttaPmZeR1uj9rNCRMBNa8pg8SXYm58zG5bZ+SYLt3RaXoDkv5fqIrOWL0H0f7ga
7T/clBavBIUz3gcuJFWuvbcCIqzOmmwieklcjVIi7Ln4EfQ4ywFUV0gRKA6Q4BgOZh9BRB4tvgvZymG0p2cESPbrVxu5hyMLDoL4
eQyveRd6PU4KDxAiQBS89a2IbygicU+mm6wh0KvRsLlNGbZs/UkSffr9xVTE2/o0omR3ILSbSoyd30tlgr6h4tFG5vnhZZR7XKqa
3ISZhQBK1yfRES9aB+zqcd2nQ0o8vE+y3dz0T8zG43nuzZe32NgUEfPZ+Nl7O0KSk7N7R/SC7zNPyMH6wp2FKissGfj7BCJC/s/U
aaqR9/5Q1keFlcCSDtijDVY5xSA5dx7nPwAeo5JZW1cl1nsKln8VXyBISz0GlnNDt07W/OtxkSmYdquYvo/RXB3DGL5poff6E0eu
tlj/iC3YpD16Jycewc9q39gOOzW/JdDwQPtCXIYQ7RZPsNEPwEOZMb0MFgwb8tnrBhf04PSHqwdTrZ6QWv0p8Li+PW4exaL7VItq
U/PjabEqFwQj17W9tpaf8WurrL7WSnUBrOoM2xBXxC5jbFXGyo5Opaxf5WwUooHzUP5zFddLXkSLGwJ1JbOlvBOK7CfMlVVjMS+8
Y8aDhxyoK5Q2o/fA8fg6P29VxXo9OsBLbKYCU4zr0wmfv1aHPw1MK3gXBGkOOJB8ynAtKJQxUQ7jMSGZal2IE3Q2GnOCVWWlr5OY
izJ5nG0X9bHQXxj4RB3Bp/8H6PRdardbQEqxq9/uy7m6rosDv3FYC85CKv+y4YzhyQdsQQlZzc/DxxT0IXLYaJJiglqXEh9h5tUg
E+XLhWqfhABuYtaCNMdA+PipXPzhpQ+WkJSMGxfWhDL7AkUWZB/U3VsrCr5hr6FRZwxsEkdQG+zn2zQEFC0CwYRfDGShpUfQzC8q
KatfkpzSSSmh/aiQhWdWAl3Fqfow8BzvXCwnz/rstd1ox7ThsZTcpjWXvb6LA4fFj4fSduEvvaAFg6Yoq2ezFJSBWWJHbq7IBDDu
YA+iRvXHRSKDaj//3FLA4Dl4sS31Xm30Uf/Txc6JUFGwelQOir9RmQvIAkDvhfXtOkqbhZt1X1synPPts5E5iAPmgax7Bb8q63O0
EoGmRMRriEfmdImZa+K+m5UOBiBkyjzWdMhRM57pZuIHw20Xo7VLPlVIEmz4f3Z787Na8L9kPnmpbCeqhakRkHn3B2Ltj6UEuvSi
e1v06w260t56IJDVUwdBpT+87rZTAeUoz7fbttaUSPKgbFsqalT0X8xTuZykIWTRphnD3TTFpaZbQaarinOH3a0j2N0x44SHHjyN
mviEo9mW1NO2Oe9wITSbd9dTk2835HrhD34ImWD93D61EhgfbTvdx0cnRZDZtsdjJoHZ2S+i6ixlHlIt2ddxE8vI4dgrhym2aIFH
cyYF9jwnYZ5ID0by2cFKLoH4yxmufnzEp/FJhpJ5gndq92LDtE/Z5kZ8BkDlmlRaXi2+glnbFYfqQFDn0Gcx5mLIT2X4uMiCBeJM
PwqK2CxHceMJthLFYm6G9iboNdB141WcrepM9eTryxw92OmEKTkWvmvsEynd3uPq9IFSq8F1HVkZwCW4f7SBFLsPpb2HyendyO2m
MUolGu1QZCfZEqwfBWzseJ3/iYCueqeWkyJpF3RqJ/wqsTomfCJa89ueeQMyz56W1nI6xNU//fL8NIazfOKT/YAJCnm57bwg8kSt
KdfHPaMJRy1C5COL+PmMIyJu+BWvMQ/H3wpQwmMgzWd0SJ9uDemqB84zefrelzWxwLhz4VZne0/xZPPfrAMwGowWmcj4Dq2MCXOS
Fi+F77ku0azq8whHOKImnXbxKQIo2MgcxMsKPvtfgNqslh6Y92gyzcjrWI5Ndsi827zVbLNDO+I6Y8lRB8C5S1iMbzDXD+Dn0dO3
eK5guRkGRUi2dueHej7Lw4ZmBeGZcCC1Ir0xQ1U24aFx3PERXAanO+chaLeYnShqMn/MDcedf9qUAu83uP9wUJWZZcTVuJOq8dNu
zwbZxh9K2rdtBExatF3z/rGrrtA5JNpdd6+WPpmU8BBMyobhd1pWKdiWDGzfKBPzB3ApQwUm8b4+7QJM4D6GdGMkGZP3jWoZ/hpP
Ix2jlzJ6hL7K7/mDvLsUg1W7yslvC+wSlolS7jmVwJkXO11lxWwD6Sx3bJaVHCfaneo+a8UVK4awyjaq6ijvSWWC44QOEUyLxk6j
nvm+8AtvC/CHYzU+iuuQ5iLJ6nPeQqA7ygBbKnYofjyNRDdAjjzh5E7FMZpepoHfVtpdqaUyQbyQmilih5YiInSHFbwAn6Z5mt52
+lIuUSgsCaicscuaE8XyUPPfxOpq7bZPAt+shDkW7DfJxm6mGOpko81xz44SsymUYdb/GhkKLiJq1G7NqiCXoKjVx/3o+olSmGvk
fWAYEUukDXzIRNdMI3wOa89hl1F9X/XDrj/X4qTpK22ACG8ZoNd0LvAr59dXnfhdLK8kTpv48mw/ofhjJ8MhNr3RJJ70fT4KywbY
86cwuEu0VA+7XC8X7OjYs6utnfoCo8LmEl3jc4fn+WXt3KYTvv6QayxpB4bnqOQZxYeW2jT1T+iOX1utOjvEQ2VO6pbsFzYlJy3c
cc3GJAcSLLaEQzLUW+S4OuLc970UD2KqcNf3oY4SLHj1Hba2m8NrN3FUMdI/O5P0ecKs0YgXJ37KdIKwAnsKUfSOWiuC3Yuj73ur
sSAxq/Jp9CRZL4q3wnHr/x6+Aatz9NXIkMHYADR8wO1WSaTAVP8WmJsin4KxmlmW1MQ8GUCDyvoClBO8oVXhqdDEvHeO0V+3fP1a
GbtR8ulDzqEJCoDvxgE++FOuzxh4KJ4AA8+gPwvA5h/YYPcpx8b2E08EFEaPRCk+APTayowTfpKjK2IGaVffDsQoCZ21YY+q1pyL
53G7rnscifG+V/sX5LGXkGp5RhJ5ZV2kFMLJc+h/DfHaKKQzZ9585z7bkSKKRLc7hDo7b+jJypSXTLfK3Ooy9DWyBC4lbDRBgtZA
VPZFW3BXEIyVfzsSk7uTflr/gpZjbjtKr6CriWCfAHC0o3PN4NPDt+LzX3YYGqbntr9jOnqhpszS4bm6jCBsKT40jJtoKGez8KjF
dHPXueGN8IaLhNezTWF7zlcLdMpvDBrusN4aqWVIC10mwM5f6waAYqHgcaXlIcGCeqmNiEQSNWFXYRoY8ov/FEV0YA1d0mbTY79s
Smv8VFwhIAzGfyxI+fTyyi3RGDzBQZaD4pH0vdl2KsYzVaQ1XKrP5BtUdZek1wmc81G6iNGalrsb1tX9xX9FXDwI3vJgb/Nk+rev
PdllVN09MlvNoI47SZ6v3ltB0PVqdF14hJdcVigpHyj42mdtIjbpPvdfSLNaZ6YxQCB9S7QbZIpL1NhRDWNZEuraUPZR3L6X1oBi
qDRTEK2ZvMwHzIWjslQsTlpaWIVRrMW2dehmeUKBABhNMIhuGZ5rHl1ttJ6+ZlEb7P55CQiUB3r2ex9LkP9/1Je/IqC6sYfJJ2b0
ywPJEtDTkbW94O8l+RM6WEWXjMW+qp8ilojrBD8Vck56GNoUDh6iRGeVc9eXCdzI0rErOOPCxV97nyvuSJ9H0Z1zQH5ZYOteveDu
Z4Uy+F3GfCFEKGMphqvNJg5Wmn6Eu1TmQy67dRZFWB5FHJLfWY3f+HkbEP0k3aGa3y0e+bgrEHxomcIWVnvSgIGORSoPLSsPzDCo
gCXjXxmpI0o0I/u8Sws9EiTk+UP7wmq1H6TfVKWSQHMLTq/sy0tbOoYWpN+Y332NYL/QXGs6id0yp39/TSZ0b0wSWs5WSh4Y7i3H
Wa/6ploFmpyxZ5D60dDM9uiqlPcN9tGxFf5ZOK+PUvuULypVj5/ih793L1r+7Sb3fgM0OPBPr1mr5M6514G2qCUqFrhp4Pl/XWZf
ZQPwUI9oul1OmPCyJmQwu8EPEuEr0nGipA4GRQl98hJ618qirBYKRD+GXKkM2rOjBF1JdI/gtn0z9bakmdwzM3W2r+36SFVcuSb/
iIlTBavuIUQS3Xfssc1vaWOOlg6z9I/25K4Uew37Uf7ZC8JAxWHZ4qbjTG6mdyRW2ex/aqF+337bVk4Vh4qg8S144hiaggcVfW/h
WPfxoI7a503oJmLv3iKsuc/rwdSdmJxxMJLuom6oosi6osQ+jktvdwr8GJkUpMJmZfTBpk0gXloaVO0q5blN8XFEt5gLOTJ8JIwL
C5LYjcWLYCzgNlbWi6MXBG4jgeqImCjoxg4WhefLimADAmGnWR8N0fHspqmB0gF3bIcju2xMcv658//NRTFaM6EdNJJ55ruveRIx
vJua3Is7YL7F7vKO2WruYCV3/o8jdFdwzJlfrT3dUywTxPLfX98Ts/TbEJGu2G11tCbQIaNbcpGapSgJFAXoi/0OxZAuRyvSdSSQ
LXJBNyKV9rmWxKOpzSbvLZsWUTPaJ4g2pbxXO7VQO+7jbVyPsSw99OE52D4ZVJUF7JdK90BEyFBVjr66FLPORleou5wYPXbgcShj
bDMxhmhVQ0kVSPW3VUBmGq9OmpBXun/sM9EvWzbZHbkH+RN4H/8cBQXaI9Zw9I65pUSFifhqfDGCOn+/pfKS1ECxWdr/ZwL29+IM
6HVaXlnc8QbekngRge5PLxWhnT4UEuuljTAMNw58lwXDASo7ms2qDHhq44B1JJFBShB1FxcyPBLP+iY2F2ymWtzxmzC74eotq+ya
IJCm7lagDOPWoXCGXLnh+aWPq6wR7p8sHVQLKKmpKZ6DEinJz0BUS9ldl64+lZjET98WMvi1G8VKzI/FDTF4R9Gve3FBpUVeTjgr
sbxiSX6sVPcEIGspz/JZmmwm5n+Rg9B+GThMXXhjLB2IblctaFs1y2tKx9K16VWW/7dq5GHuWe+o1DyuHKj2KWOTdvBJLJkI7tTm
qOaGEOJB7amkGJBeEMs36xyTCnFF6lbwElDifk0k/Iy8E8VOzXplyKWKU1bGEAMR+ULoZU/KqXHusdtu0lXwiAlqzGzxKk6Sd7ky
/3PyQdpYhjCKryGpxidrJo1qTLPiHDDOTpYyPGfUPhR3gvJkjr8ybHCBXiQkTvzS6j0Zg50H5V8e37U7ln5oUqeR4qU8F7FQ7un5
6iM6oVppVDW98GQh950M7YTbtbCFTKQzLmxk3vYwKKubYV0bSyOrYObFHu+69SDmv/efDAGSI+bfizBV50gX560qa2QhF1CgR5zM
y9tBIR9IMlSAyDILZSKN51ZfA4QF/Wk0ylkwqDbbRKwXgEObgLJOhG+0JGwESnnbcVJXOAXo8dIDM8G7mrgH5EuLz/KSpql0OUzE
xkAUbGUMP2/VVbCi248gNYDvRjBsZzVL9Dx/hrV+dYaEdMM0c0s8IBBwANiF2fe88k8dWj/KHbT/Dvmdnp0Fy3OneZ/6APl7fh0T
aLJTCriqqx2yO9ZnsJt6CbcbKi2fRiOTm41qONCtvW2sPjunxkrZlW0Vs9pPV8S5cj9mV6vAytHQ3O6s2BccYeSAuXVTiwYkACbv
RQu2Lpc7COKDF80PhvkBSi6eFjXxoDZa2lgFD9noLoTyRDbj0ISGVMKkBOjBj7GPRvlXrm7IHtZxCjtWR+HyjOGJ1Xybr3cCT2nl
sgzYbDTMzV8auNrjeT2bKqcctppxYOZ971JnXcrKXK5wNi7il/8hzn6DAylbtKEVamSEifTPGhDXCEbDT4GDzR9UXpszPD5LzSWj
nDAqLhvyIvTPjqfQW9lCYYY/EtLcabV2PieBlDEg9LVK/QwtUXxcLDOZJvCWAQBHtoyRv/CKK45OxSPTqFyUGqreP6gBvkwQ84E5
ZirjqOMnMJwHhJ5UjQx69DmsYpYkBuVVap9NWhZDYYrgZzqG+YjWKQDuSzkDJtsgbPbqvYSMsrrnzMvYyNDBh9VuQYjfhTBuBZZb
+RGI6PpJL2sjJHGsgt0VtQRC6weYbVh10jyDGqAaQCIpucIn2nNfxlreifTX4S5Yrc+fhqCBAcfxUJdG0lRMxfrIG0sSDewEL2EV
2E3Vwbr8viLSVrxwZNcodQuXGja3kHAt1XwnTVEOT/g2ytkj7FtXBqQPH+aPqRnz2pLqCOjSlNCX4yp99rjatGhCOIpGROqyw5OD
ZjvdiLFCxdrpej2vEQGh1gNis/dkKAJPpdpUTnyhJpCNG4pH7Lo9mA3sRCJ1TEY+CpV61N4b7KyjkWZolv3E1cHiciw0FYmn5bLa
i7k7JlH8d7inkEeVBJlJap5I87vtf4h8Jadl/0FRUCjK//5B6YnambgLO2kZmmnQNwc1GPVOpLz4KqBHumdgUBYIhEIsUzSpSYrH
PU4yIs3Hsdymleru/CRYOQpxZXafoNWow/+IKKDhVBfTcNxdBWn6T2bek/uYZb6gHPbLIv9NjRyiqhFMvPnwipSizypVLjzHqhgW
NrMfbf6aDY7mV9uLXO+XdsQuGRo8HEDy5tjqx/pXoXYI/sNWizSVty6uME47EiFVGIEkspAfFI2+hOKka6dgmp1QbN/FsFxj/tYQ
ZmMYB55ty5arU/JJGdph6HDo2fT2SF8WNmM4rPK7AbFZKHuj4tLhY+COk5K2Lydyzpqd0z/pXxU9sgfJ4byTuNC82rLvxxo9PoX1
iYlxUBAb9SgAfd/hqj6n9LffOMhFGxwGb0Pzr0b75YoUbvsuj2p0MupBBKBb27jtpsoPKKsB4WUItxghZDQNSq0QbbkcHsvJNAuA
c7nhBUvFAaksMLrRxmcDijJ83BiWvSpA+nszqItrnxMsRHU9eas5SlLB8cW0d8qXvcKPZdtC/c8+w5/QDr3KK4C+rNtpTKpvGTfQ
XJAyDut45jukyebvMAk7c9dUKm5u1CP0f5bX90wm674XAxir1v55QQsfoBq5rKr+YUD7ezRL4tKFwGCfzYZfwjQmw3v9PbaGXq9g
9qaI5Ssdmzxiundc0f+IolkMtjfPw0HcsmE6ggDkcVNiYgrUatEZNw+qaINMrXLh/QsScRLTzIBS0d/tqh212pLiU+lnjMuAMuUI
HSmlTfG/9n3zVevA84AA0LSoM9FjzMNOPOOfithqWvgMP0fW3EP80/1rODQfaJ/Sj8pZjx486Hd9WSXLk7pn+pzCwRGPN1YnWudF
fAz75FLE1e/SRfkqRFHJWvI4+yvUWDufLDbOoUPctqdAjtKVRnbFHvpvpqo6LjDrqKQ/WGl97oRj9nLQZFZK3kW4ujN950TlrL4P
Mr6YxnDMSFn/uksqr8wdhC2uDQ7GcPWZlZWZ+6UgtUHPrb8k0doXcI7Ie7Ivc298q1UDV6SNJsc+AmfR3AabWcVjWOXauqielg/5
p7tCBp0ZmwDyW57zDTxPkr/joLpew2CWft+W758RL5sb0ZS4bByHdCicfzbQkWK8vFmrgkvnaVGVfNAFFJGwlrUkYF3ACbRvOV6b
frfsGpmMqnAUoJOEIVc71Gu5p8zJeOwoypMHR2c5URsnK9TlU316lN6Q6Ut1IJHPzWAG01hBNhVMzRz6znP3Vsi47wJDW/rTtrm3
ABb26iHbFa11wDpFg5bLXE5QzqnQ3b0O6pHmI3ZPxp6JOhdOM/EM6bFrvEW4BREazG25niEdOyyoVpD6VQNjgDHpv54HP8pfZOtm
Opsta8F9kzq92FWGnMmxqjdG1WHoQFzuzrdE+6t3uxgj8ffKzGpjDLhTAeKa8fFIL3hYSovnlDquEf/ksCoz777cfzhp/heJl548
PVpHoApSxirIjtLyeVSRZDMKVxD/4xCoyiz5Sa6m8DBr5dzZge54k7igGGR9CEHqq5hkOGYpehZzQ1Mmhvv/nsdDvcVAuUBhyVrZ
k7gVqLIU+HZplZvNEhUhN1vWVJtNauoqG+qcs4Nf4ojo/G8wme5+e80CZz9xErdyQc3PTEfR+o9TQ354ohHlJQjhRnd2DxQ/lkni
OZ33qCzcWk0TuXBQKxX2s3Bm8BJklszTMbbgRSEK+P3HyLxBR5ZCCD4cgJ/hMv+Y6l+6e6tf0JzDiI3bXzFoWAj16WrIuO7pqn9q
kM0IMTnZI9HdsxlqDhltaNPzcRLEQ+Har06GVAVZmd494eo5z8GcMRK45tF8I16mJuFq5GVT2M9ObybAdHtX6VO6o74iItnBJ/+K
g1ZAONjaYN3xmBgiBbp2yYiuoMw1sT4RE/8s1LIBoKnVe3NynXErvuJBQwOnMa9nlMnon88qLTJo4LEbBLyOdjbR3ppvj+YwOB24
OYxpQuBOe9jJxmdm2+io5BKXZokiT6I+Kztli5DIYO4AqF+liZ5Be1MlkTd1eLpcaH18sVX47IU91O8TanZkXtauaIZLmbFFrOgn
qGI0cetwepGaM8Cs/Ovqbfvow8UwppnwtqJpjDSPj2KIYRYY3tPcricAtIx9zSbpIK0NXMP+5/KkUR5mseN2SllgdoqSJolD6yWh
3rTGzYuEwVm2+VXwVJk+xOauRq3Rjai+skGQvZAt2vg4bFYHgQp+KBVZbwVmsgGRDMluXfnm1FlDCJV/wuqvvJ2uz9vcsWF03VdQ
+cb1fIJgl8bq0tUDbCOKcEQr67o2cnW4oCe13SIKAfaIbwzU4sGJJS26LdXR6nLePhU3k35KT9D+05+o2hX0KHWWkImSjVNErPgd
EHAuYvSV/niUfiPCjngQPMkn49Fy3J/wo1LO54XLVLxlwDK62vYi5yAufUckZwtZ18mwZEn258KYoie05gn7E+WeyBucX3+vWawx
LfiWegmbPyhpr0KH+dcrKsj9+VxrM2HesZxlL5HRg5HdS35NlB9fdjpr5XWqOUUSe/Ac8pcsnw96HrBVDu2FMgMIhzQCmO0dw3wq
DkH2gi+DtoBbpz9ZQxeW25dgTyM8LWYHjfwBldH12OwTgS0y8Puj2KZxNzy7EQ89x5EBgr9YkClS5VMteceC6fKUWxC0OSaT1xpo
eaXboJEhMzrKHv/xGdxbjd48H/6BCeERfGlEJ44lSReteYQfqLO7FMhdTVapeiu7x0UPgCMdCqkEZMGdl8c333v0D7pnV6DEDgv0
Yk6mukMSPm9L0uGKVLED4XvLf3zmsqyQ/iaYtHaUuB//ioRY0V04gR1NjaW3CkhqJY4ADUBl/ZJx/HMBX4pMekYOMxk/sCOiKsQc
JQjwGTPk4eQDoXDyIy1zrVC9uz++FiFT0s9x5Z/OmNXlSptq2z7kwMEMImEcZt0kjWvbeht9QSQzay3eke/7mr0XFJiQu/msja1I
953JOikowQ02F/Lc4kksaQ3y0tbXVrXsJi5f3Qcy0gGHw1M4wtngRMV9zzlRxz3IzgqCRgZ1iPlru9VkTpm4/cLV+cvGvhrovcBM
lkdEypl9eJuuhsZKa/avhZGStxn5T4PsYsp4F2RVJCoLiJ/bqNSLRm6Kb9ZqMcW5Vp+TJIwnhgRkxC9nvuPTSabLFXMwok30IV3s
FZJ4+UoCV1lPZHWtfXd2i0ruvYLnMgqy/d4ofWEmpkzHxH/OLoJF8lP7N8lQlRq3DOx9kBF2gnaxB0T5zNkgI5WE2/cudkyRe1qU
3o7JXSYMC+EBUC8yeVuoPclnYIoKxjfeEF3mC3duGmCNCVFwABTwOCAvp/MVirg+JeGOMv1xQZBcfZ56MiKuRVlOWE0jhDtsQ7IP
Io//JUGHgqODVnlNi11Tyy6OWufkQP5K2oVVR/OkF67TBUpn/6WC80zfq0lPOkyWiSIwSgurIQ2abE3I32rgQGO8juewKXjoSNWL
J6HH2blYbn6pgBhl8rhm+/Tz/kADGM4SnMwGG6OWntryupCBOW4E56C9Oy8aLx7VnzEkwQv28RFLINrQNnJQX81AeXhY5V2wZNzF
P2Bz1ZYUToT/xz6H7CnrYuoz+DPe8Vm6CsiJbVTWOoImmfLC+EooCKfMR8Gp4HJUak33owPVhtoMQhi1KuWbllFk2hheaPNcXO/V
ykvBv11kmdeinKUn2wBelPiMJcWN97yjSjVHbT7+GMamzQ70MMK/aI+zTO8EHPrNV+HA7v70WNI6zQbbyy/9GLb/zvTfKQM/v4gb
jyFijy8avowonoVBLYZedoN6NPgls7nbgnrUkFCvvl358KwlY9lXBqvTrc1/v+hNYFvt6xA2OX9qYTLrMVPtThz68e/ieNekIuYz
GX+m6KzZgsV+pe8Oblil5TxFS7ZEHI2idf0E312+vHCGhq3Ey8V7+8mA/05aoM4/KKamIsS9bM0vW2XlYH82MZY1HOqu9ePlN+0z
iiPWY5hrIHqn56iDqrZd65PenhutZWF4n72a2H+3NSH3J7IfEfVxWyL6oT0+XhYRM58Aa05XZfi3ISws0a6Sesf+rxYx3lEdUW4p
tKnOIATEZMjx9TkQdeLUj0fxOg/y72gInTA6PToVTAYABWSx91gi7fVIevCUfjdSNXFFOElzlOdS9icM+BbDBHgnQcobIX5SS7Mz
bAhcd259lJvUnYRTSQTGnM4WFTrDeG+wuA7XyWzH4D0oW6zjuZFaSIOOZA1ACYLXncpsSAvCnwAXemnj5FiYVFPZOiCNGzotbC5z
+z0HHo/zoS6UtzubQE83xLKE+hqUMVo1S/V2wnk+awi4uidLLaA7NCEKvsD61yuBxbd/EvafBa/YzuUrUdJkcX/uekWNmwQ+l3A/
nIdInouQEuQnptlkgRYoaNiDJ5i/MI1Zh1by7KrpG2RIgfIFuImhFu7tMZDCUDSp4cSYTHEPFmWl8iCsml4RC+fpQa6VjQMMpGUF
a5wqfNRuqzTUpwx+D7+a5WaXC028H/egHuKmZWDA6i/mcEtefNsjYs8bc9N2B33W2GOeFnqyzz4rKwXh8Gs35Li1jtQRvVXXruP4
uWzp0GkuHd7FRXQjt9VRVK5qknfV4nJSJzYjN+Lrdh9CWZD1ewboMIO41utKzzRsXveRgiwYDXhVbvDD9uMC5P3zuNI83q/RK5D1
PXp5NB6XS0yQyACAp2hFAMmpA21aNefyj6Wi+jWnArV6KFhNYhw9IFHUZPBL68WiYt9nb/ar88fgxPDHIgXnOfIs4STYYL5+aGEz
fpLoXIGqY+u8yENFnUFHkOanYrfwphDk6O7QxUyTX9ez7vXujPiWDz+6y2022TxbqMdUB0NmuhqKWs3MTBAFBCqjXcAIhbo19sP4
HLGiX45/L2VJgXU+TK25iIQLNSEnPI4MIyWl+5kWf2T1Z3V6tdFNXrTRgPj3wx2scU4d+sm54cd6M0HmG8cgMBiuK5z/NGCzC5/G
VxhtAsE0ImSC2KU2xb60RJ2ySsBjkQjPhzmZggmkJ9hu1ISyWw3DJfKOPniwcexYIY7KC8fg8sZ0wAC3hZd4GelEY29A4lpkYtcL
vCsBzrYzDSKoJOVtRJEXUazvo+5jfSkYRnOtJizdtmCrwwcjyGZKEY73PjNJk9QSpUdgBgrxRGTeWU8FAi07TIupMwp87LbjoZ7o
oX49PALf3LoSyVOu+7GD+HvomDSX/K2BWVga6S3VnENSA9T5Jv0nzasldKdXTozqkU+4AOrRs6eWxHxhO4lOV/SrBwdlfHGTyDK3
eALPy1+hMujbbaEpkkHIT1VR1K4SGVy09l6mwLd7Dpkkd1+PuEfSvsnkGToh9kfb6i0xaOkvXz49vTSZsws4/QOn4bL7r5QBoGFs
3epwEdt/Ki/SbHiFVms6h649ZxZCqArb0Qah1L0lNI/YWZbfS+eCYsQNZVIaBpi4cPl6vRgRFyvOq8w7D8w4vr/Xx5CusiMjoT4x
/1COAQKp5c6TZsXT3+m0X6avWnJXYEHulCnKjWf2tCRgiszHAqOBU85fkyblWbooJtlseQ7ieQ2fC6jEf8rzDmjveDalWz5MAiMH
Kl7BGiBSq/vojTTC39DQMoiASt6apb2Ht3lNCNagoLCYWpOS9HdETO/VsST/XD3zIwE4Pg46gHiOHypOX52BYjSbKtvfa58vb0tE
KvXwbvC7qZgptD4Nr4GxeLFqeZYtUy4ziZtC/wOpkp1Glm/Eg9klbT2eHGF1xrB94h2a8/qRAka4lFuLM6x6nnxfF4CT2Q1GsSC1
6VXUFGyJKfN7pT0N3TUolV1DXxoXzHWa3y9HQOzGQ4GgeKv3gUh6HSZfLTr0QqNWKUMo4yMbcknDSCSsVL5t4S5dCOrsM2wcvVYC
TeEVKPHL/VkEa2yTZjNR4qVstjWhRTHZn/EWbCJgUzeoQrGonjn1LggKq+cejfzh/LmTWjKGNoi0LlGJmOcjL+bFTf6eyHagqlZi
09KO+Stp4yRqrMuifA8eMQwZhxwG9wsSb0NsHlhlWFQOUeB2EJUjJyMh/aJhrch/H7GVcOCHG2TVgoFGuKrnAE3hH0Rk+OwMInW1
xGXD8B/hIJcj/bZoQGA7XLXPBV6V2Jr57GbE9XSBYEPWUrq5QB7oFKL0bgIC3NwztRSxAJ90QMnTLa6xDO3iigRFPvYrJbP3RVug
Qh8+7ftYTquhhjYp1f2pYrGqGJOqG5hLiQbGeBvE1ecvwQFKFUzByNrqE1jARHiHAe8WVS3QRgwXQIdLHi1Jf3k+uSkYgp+hWlN6
MPenspxcfGqw4Osh5h+vdc4K8tht89MOLolhsHPXeBBJTmneiIZlpUV/hOzOIJLF6rJxIOD0+S3hrSK8HwzE/jqx4bkpLkiN3U97
AfIOe0Fh8yg3rQCNrBaVyJ2gApkDsbeM3DyLXits+EHH8BLpQvx9E8FgBFqa7k7cCjBXXDS5PuYb6wBi9yZ0xM8yO5fG7gvtZflC
EZxnfAIT1Z2JPQ93sZNcOahyyD5EXIssMPZ2WnISV+Ycdr3ul8DHl1agDWukQhg9JIChr+Wyhu2Tu+KtzUdtoeA3sNHeEjGM1bsQ
POqIXws6B8RrjPw3LyCNgIiDCOaSZtv0w6fH3IKOV03ElWt5oMaqRchhacl/oLrdZX0yDXOS23Z4y+1m0eqjsBTJjdzaXcGBreTm
w6Yl4vE32t64wgRbeROujBk/N7X8p9KcEFUt95GvRzJKKBSnvfP+x1IUQ2ya3C8NmvbSaUylQ8qjvpvCCHiaZRyB+lQQCYH7fq4z
doH3GmPNycX8SiXf7trNH+poxqhUF2ql0+fVNI+jXNZWF4dGScZjZsh1SdQca1BnzR/DClqtmAgInayRb5xR1q1L1N+GBD7t1pgs
O2WR9iAdgMViebdKMwHEgONNUR2+cDKoqqI26rGoc8tQrSe42ABFdaJpRcAQrVnFbEQyssb+yqXZlllphmKaWT4VyybH45ekMg9R
2rgJSMqhRuamCHA72Z5mHcP2+DU1iXpWMWQ6i+pyWKauh9GYRaR7HxwjYnb5Yh5dwJTnFZaqP/oNvXrACeqVDH4r2GsRQRyrYrBp
Rlew1Dh1UkkQv0cTng0DaKKO+7DrwhvIwYaAMdBSsrO8GjyIDhuWIka1DNkhEs4Gl0R6f0ePL/fLdg+VHWpAP5e4woHcW2PZ4fKV
no4uWczDupwemzXnyDoZNjIiisrct03zSxzDzcZakSWuyj5uinxu3j3U7sqfvk/W4lvOb7tLUldJ+cGM++N8z3VdVrzaKAz9LLF6
oA3lcGq3xpaYgdynmWReq25scep21sDghxuQpksYLsrQtOHwqQBiehvJxj5JXUW6+Ok+XUWn0oC7BUHudQVdTQA+oOpgUnE5FIpZ
9ud3mRZZzy40TeQ1NtoLJ9yfYAWefst8cqvICbKEY1ysth5eyy4Ma/c85PHj6Y+MpdXbOoc1ZYADtMn2ZOq9Ywh4Znzqs379i8mZ
i6ut9uSV9nyDo2PuEL+NQfKxCyoipFuDsGCO5hbom6s9Srt3vaAUPrp6G4PEXRSRDEK0b1jtTQ1wzD/et26a7UtBSi6azYop/eJl
lGWdHUXY2Skjgka1t3fqEGe7WgF6Qgx3MMhvdCdcnUnaVqudRSxV0JsDqwyM7fRNylENLk3pcbWQahIrhmPx/hWPi6Bh/h8NuszC
C4kDqD/KH0Z8E/NeROJbmdk00MsvBNEvMZxFFlUvKjIhh7j94ETWw77ydwXT5TM8CQgdPUeeLMdsDyUxsJARUpuKyEBN67rcDwZM
vMzVBy5WWwTreWvy2jvXYXscZQmsmwPSOuvoEtNgGye76gqIZI5ccVDNDsMO10HSs++YwbqHCtG1L6UiBlepPUv71epETZJ4QObd
OUMNLKnwKT2LxxXFVkPqhCNIBggjgUr+5RCZbKmkRB/sMO6YKPbj/o3vV+IAN95Nv/VkPWOa1O9txqlehE6R7mtRwHREsJjsXl0Z
BJ/5KKKP2CuF3kDDsevvWK/t2sOBwsKO0KYXRSHHeh8n+7oTnAaPUUFNDhLjs8L/6gX9i+e1FE47lPgtDJdOUMhMfsmEing4r3Eb
GnfaK5X77L+TH9hYKFcKdnqygifbi3RCikwRaE7/vqK7i964oDGjle70iiuzwn4G449X4aukOVFAxtib351NxNXHypt9RgCvQtL9
h3PPsGATp6G9R+ZV1y7T3OoCcDhsSwpv7HQA69ySt02C5r3t8qgBRrItetVS8DX9GjSfBcgvPVkEGh9RRJeJCmlo0H7dQ2SIDBLh
/NCGvtNxzYbKTn2RF6HnbTGvp4f3JQ9Xll2zIF+auJ2xhF15Z4MH0IQKXMBA7IlqhukDViU8tDl8Qw4YBq7i+Owkbe75B0lAM8HJ
PAh71GZrovdxlAbUGrnLbfgJtERYE/rOlcczxUucbcxOYvxzaSyTNygMwHplRRa9SCPssbHUE94s8DvyxUW8Qz7V8M2B+Y0zKZpo
qqP5Bw6qIUEp6mQri97fAaJ595hxRfB7a1rucBqYZzO6jQaSgSX8bgryw5xDEBNJmZ+DvoDl3HGqWxcjpy4Zo8P1fPVD5BKpT9C3
mLfY+1y+wOlK7i2Fya+DShryfUTzGSNqh44W56mbmXTiLIex36/q1kg8HNtNS/uEvm7eRoyR2u4/+qISQBouhKQgKZfiGmdW2+1B
ndz9wpmJZfyGncuVLxxzakb/eJ+g3aAKLsvx81myMYGLvMdRVXoeS0eN0KUs/ycX4H9s/+bODCqeB2JyOgSEsAIETAxcto2z0rkN
jot5u/CyQSz62Lf/YaPc2TUxTF0LsKHTvr7ZDpGlMxX7Gu9z9NNJ1MisNPGVUhOnIYnALvAVu5KYOyYWezg3MywqEXsZ0cDKd4Ww
7sl/KBslO4Sna8J+p22eMUEodsNG75VfV0rypU+3bBhKbw76OssBMc8IIezrZHFb8IzXETi4dl3Wpkx9LATC4LbUnLG7glrC6hTo
HLu18lWeigePfP3/zk9M5WAteoLLIZStXZtYCVL9HdT9Hz3L+pLJkgVKDJtfiOpDh/ejqaqLa7Yvzo+so6tBHttckBW+3lmTgKDl
Ox9N0TRec706p9Kg2ITyTZgH4qQ3Dak9UkiT3cT89QqVqLzNTuXfwR2PHpplxngCLjBcrdF7f9WVjS7UhjdfZ9RwkUCzwBYhcGMv
zXIWOhROKEHPkrrDrroABX5SNMEIMgkSJC3xyMhxnNEndyCTvOzS0KkYElBWOUIvC2h3sY/f1uMGO6Vb8ySoWuiZz/4ZtpyRJ3Hl
xfzfx/6+v99njV1flj/fm6HVgIGS0KEyhyOPc3RTG/7iXrQF4MMciuYvMu+pvXc5G+ik4N1bUgsKwXoqA0tVMMyhiB9boXDDludd
5lReo+b/BEu4tkhgz+wYe7LPjvDfs9+5K6da8kwR9LkCGX1D0HKPucxSxdDGf6fTGaIb+7ECy7scQiYAR6TjWGXH0VGKFwFgxJIC
U5BE/Sdp+qkPi658FXqsekOKH03xIqSWMZf8km07F2Pr7SQ+yKGzZkNJZgizvvVsALLnihS8THtiJ5VKBBjzhBLkHvGQcRWRKdzr
UEs7ia2YWQf56iTwzRDQQlwqfY4tsEHt5FN2QIXlmpvwOZbKfdj+UTUMaS4X0S+mFr4meSATSnoUUm/FYF33KGkI1LmCC0rwFwAA
JqZ36u2Jjw4K+YAiDMHZGy8ag6o53CHn9gE03mg9lF71hys1mauqOxkVS/352bbksoLd51aPEOd8UsjxXnmrbDU/r/ZWpC9BJ6lv
i2temszj73I4uq7IUn7BajUHy398kRCpipGGEUl0KsqvnHlukLO4gPgHER03RICx1gZZQp+pO5SYqNtzTCvr+NvTMdy1FxWgQLvF
o6qnQziPOpm1dE27NbU2+MbQMlOz2hBNxEM8E5bclYmGfZ95J3I01JyoCEMwyc75n7BDHyzgA1mKYNckiAIsU8/OJE03pJgtj4Lm
e2CyLL7m+SJfEt32OX+WstsubeCwHqJFPQP2tmb1FVEYQkSasNubssmw5J/ODpNXLrsE8iU+hSTun6sR7HKsKiBCf/dUfSL0XjRw
Ydpw2xHcJjJ/AkfQVZxL/OcBzCeaGgIsRRoF3ber2++oq8ylN/T0LJP7SlQhanYk/L9m+cwUTer6bfCW2wAfhwSw+/Y4JYISnMqx
kW0yQFD/y0bI2sDauhnIwXO+mTAXTiM1A2EwrCMvTgA9uqPkSzQ+FFUERkET4Afeo3VlM0TRexB6e3JdSD9YoH+Iss26SCaDN2G7
KmJYeDLP0ujdb21rC1WBouD9RVrRotQ/JcigelCNnbZIdIl2QHtNLp++eFbSLGCBwQwIXDsbb7G+LkjBUTPr2y4G/7MgfwHgRx3S
mw5cC+ejPC8GaEiC6smtdUv/F/tOUw6MKD0/KV1q6gbrc9c1srYdWVGmh09IfuciCmImXBvb4Nss5p6jib9qqzONk+syzqPd4FDy
Sp2o1udIhOTEKVnC+UT7JsWByFQxE8fIOLYxVRTY3H3uTupGth3pp4kIZAkaKC9EcOM9U00/zWLowgJuYo/kE7+bX6agyyhy9/N1
F2f5iRmcYPScdSWn8CsYsAGJcsxX4tMPb9U7IfKHnRgcTF5lSvuqwkwpvFBQyHpf+aJuK7kGmsuEK5HOGxPQUTjDWSx0AMlZb86l
5FvAjhuBxndORTZGDFORQ9Uv+RkFb64cSGS3QkuApS6YjdZoc/W7dgrICageR76XsC7kJTIADilPonkXyGb296N3VEMTNaVzGHcY
1Lh/pO4Prr6DUWGujdndUGZ1rBuZuqCi6K+F9jYQ2w0wnMfWOlKsUDxPP8uLxvXKSaFBdxAfnLOmlEXB1OOcSrc8lzhebdmS1hRp
aiWBkacP3HHKfKcDvBc6KLZ6CZHw/urmV+DTGPZp2/PvIsLEn5tiTjINXKE7QfdJk35cIlCLDwmSJEh4XPJV8iX40V+Q46dw89f2
JYF8FlHjfHNfj/q5wt+S+pKmZKxRCOZb3n2NDWBLfYYTLNqm7hnHFm8BESWpxfIO8N1qIFLOZUpGRogkwRY8+4A+8UE71B147yIU
LBd+/c/I1LreSmbxvoB/ZlaJGjIzqW9J78rll8AmDEGpsfDCug+b+wP3XOo1TNuwjc06hhn4SP5ohiA4BBsOZ+ZaMWoEIBfrgL4/
PhJ5nprF1jzHzf03RQIrQaFkI/QZD76ekPa7oN9ZN2oSPTHhfJvbA1dYradY0/x18qGRoiqjUmlmcziJcr7asR9SObHlvUx/UMJS
TUqCY1bE/UwfGHl6eloxJTLTjJ4dGLcxIsQM1Jpykx8bSkRkgXU54zvrN6BPpR+htmhBFdUav4TGZ9Sub1mRjxDC97rhjmQsUP6C
T1bhb4ErRl1lcsHiFYzV0dA2XT+wLFDUaDhh7t5JK9Y9Yl/e8LacChlXezDcS+JXHt+A36Kq6tWf/c87ZqoQlux5STiBkqf+SiF6
dDyjqh18xZ89/4TJrojihIzPHkziWWezjoBiD8WDx1ETxrSQOSVNbTluIuVVRG0iPOG4mgKj9rrhkhGvuHYm9/ercy4vh81jM6O7
y3JQn0kXa9WM3oKHbwp3QxqenuDr1nelNFZSubqHdUZn0NDustCdQkPKLYtIYCmdZOHukCaiL7NqDJOllQk6gCJKjI9uL8QjgMKF
a/jrLzxASTmHjVPutgJ09/asmubrQdw9li4eJAo5RjgPxkOLQLFH2r7mJasMHpxKoNVWoE1+L1xj6lHmATxTrIefbm3wLygKTqt2
rW89Kzass80eN4mqOEJ1qQ+IDh8AX9TsZWu0/WCwbiQpNnvdpM/JhsDtqfHMuwEIAHeTeCiVlvvQRNs7cuG+HYWNGc2YOFpWLx2c
uAaCJJZAHl4Ko99F7uHtfWULYUgQiSlnuh+cmdcNaESpN5ekNm+pZT9aIPaWGBtpT0ytYB6KWb/yo5KpBjW/JEBkSrBGjOy7z8Vx
i5JA1ygVwECjADO9Br/zMwkySyMJ8K9nAoddKz2gAY92IIDPXviNwy3b7K1/VEdA/DuWrriIFCfWt501Ql5ZSwAJlj8p8iCaQD9+
9zlvBwL/YHnYoIM/9rgI8QiabWPVV6LKe9STMR4TWiNWzsx9pXWwSbdfzKWgIb1Qt57bAoaILzkXFkBgYweQgGX8rwj0CcB6hLyt
dtaKr9XV5EczJrEOvabVO74SkvC7MQqOcaw4p9SBklUN/ZKVpY29w3QuWCW+4JDx/m8UtJuerqOe9cimzsiC5eQ46xofMmB+l9Ih
01HDWzsecgxAwWTpefVdclWB9qs0qO06icmm+qNRaVwMp80ve/7Mxk4fppmUqGA4eYMOaWuij+fQFieAPTR8nY2HakMul9pyvM81
XJdfwFYkKTN1rfnr8vkrv9yxkxPPRHeNzKv4KF1Xx1DI0h8rjoWLD5manSxDOFQbHP4ygXo0WFZbIU1feZqOI2TqRv2SB9hYbG5i
4J8SJBikN2wK1qaEJS3HqCjboiMmgjGq8DcsK8WntYCME6YKATizj8ibSmz1AVKQSeWcjvMdZ46wCQKiL4+U2QqGlzDhiPthcHOs
5nBdBriLzyju1JOraA4vlEvnvDMmEgdctsMdzmM47Sb/kJgR9vsI9Z0p0SXwID1p37OqG7dQkWFWD3y6OjhmJJpm6+hr72/ePPMX
S1lkWVLhsQ7mQeUVLSjziXj+efTAlKoKyHC6C69aKtF+iVoqUfcIlMaI1HdQbtlXXjKLYhJM0ir+1FEXd4kqFuTlzRtqrO8akKmL
wS8iwMLL55E4hZwwLueACj+D7e+DBGqckWx+NY45t+qtomI41bL4bVjlKUaVL+NaUOFoJN6MmgNTD33FQDsK7eSjnsDURODM3/K+
CPNOxk+NO8OzH18Lg7vRSyInKlHWWmc9k+tm57+ihWXBFW283WSkbpmsFiXml/XDa9I1X/j9zzYKrt8t6OVvl9zkcSbptZEQETxr
fU4ix+6qq4ZVwXKvnU4kShe3qAncqRLjkql1hFlk5t3XRNctRzyxEXyMSl87lkLyHitYAQCmqLBkCbK97O2TTOepo6EYgbDeLCXE
IewpPJFg/i3wXzc4J3HIOnyCdrwsSHxErcNntfvoLiVjzb6II1s/uNObVLsvbX7momjGY4OyrpOBzuMljiMBAiiR1Z1fypZyL5DW
KPix8SjEp0j8rX0kzviTMfXnBGjKi6onJZoGX8i3ISBm8EpnqZPpqDmmSmO6VBwcgoLj25A62E7xc9Wg2j8aBKtpNrwiuZ5jMMTK
3erha54fH7pQt4ees5U/4tbfiKm40K0M2BKN0MWpIPzjv1DmTaCKGPW13OrqdLEzbwQe6B6+3iUSRyg4bViDQFW9Yc5kmkUkN7DK
856UV59KqH+u/GpENDVWpeAR29exAl5alb1queD3kxXKTzy5JP7u4+DxZy+xjRitf6PYWEdqe3VyJrGC6PA49KZiXTXEz/Mzpmz1
M4o4KEu37AnOeufCEZQwaTsV01N/GhLPaNWwTqXWEliNJPXftnAhxuafjZecWXBq+0m2o/wmhiDtorNFw4YCRKyOMgHz5yyQBXxL
mN+XnlcrzOOf9O7AmNAI1Gvr6AnalXM433SA1tu9DPY8eqyJMEx1O4VjSfkOndrYoR0op26EaeqPRwBJxZxWxwg7F8EIHhPG/QMJ
9peH0h1TiwHUVhqUt8S+FzXP4qcfHFdJE5t/h4+tsJ8vHRLRdB3cv9dKqyOXDciByBvsda+k1JbY53gKPllOu3Lz75Wi0vX/y0f5
1HjnDpbasp8jjYtDV1ullG2Yr6lSY28Scpq3Gf7sMQXhK2EIT6DUJqUXbtMPxwyD9doh8fTso2XBOs8Urq2b98pPT4Wi33wPTirr
M5+SMg9HT/hey/A96+3/e6fLeusEkKj2G8L6B7P6Z6ppuSV9BamewSntWfOWoPLbn5Bxi3GzJSirQLkpdrgfY9rB4T1EtbRK8vsn
slHGOQ5Ls+EFTQ6rCm1JPhLCrjdTVjvqdB12aJwvB6bkdc1nbBbiRNkmMjEckjAr9Sk40YyWglPhfOUiE6fJPbNqOKjC9jDZ9h7A
LAWe2qOvI0MZ8RZg30A4FQRsxlkiSfLU7DuHfN9tnQIG+GO7jdPdUoO3PkAj1W90pdhOK/GOwYOtPX4Of//kD98QF0MToiZFgdy3
4q4cWOnlQb7QeO1x7W6NPzpo4nK6pceKTigjRxFbfX0BNOvl80MnIEBH5Ju9GJPNZ8Ce0yPvb62ObOGE6QfvCkpyluRkrunUXXfL
aDSII1y+7nFPX9Euk9HRoXVyI+tReefGhejg8MDY8mB/vVhzuTK5hJUDMndNUijdLKK/WWIuGthjtPwGxCIfP7cyqBUnxeKX9cQM
O+2AW6/GshhtzBaHjEQ19SdJfpdrMkAGqGXrDjKDQcpCEa8jwZzNUSkxz6TXQDXb7w3QOpWs5J4uvHnIysDXox8jFCqpEhGHjXbi
LofoCb+/AFezZkXoC4eYfsj2kwSc6mzWYj8T15c6VXTMlOvLdfOBGPNJcgu2dNJ6hZTS8QMAGjOvUHQwbK6Tpsw62wAK/GmE8Gvn
f15lN/YoSeC0