Correct a couple errors in the xml markup.

[apache] / docs / manual / logs.xml

<?xml version="1.0" encoding="UTF-8" ?>
<!DOCTYPE manualpage SYSTEM "./style/manualpage.dtd">
<?xml-stylesheet type="text/xsl" href="./style/manual.en.xsl"?>

<manualpage>
  <relativepath href="."/>

  <title>Log Files</title>

  <summary>
    <p>In order to effectively manage a web server, it is necessary
    to get feedback about the activity and performance of the
    server as well as any problems that may be occuring. The Apache
    HTTP Server provides very comprehensive and flexible logging
    capabilities. This document describes how to configure its
    logging capabilities, and how to understand what the logs
    contain.</p>
  </summary>

  <section id="security">
    <title>Security Warning</title>
    
    <p>Anyone who can write to the directory where Apache is
    writing a log file can almost certainly gain access to the uid
    that the server is started as, which is normally root. Do
    <em>NOT</em> give people write access to the directory the logs
    are stored in without being aware of the consequences; see the
    <a href="misc/security_tips.html">security tips</a> document
    for details.</p>

    <p>In addition, log files may contain information supplied
    directly by the client, without escaping. Therefore, it is
    possible for malicious clients to insert control-characters in
    the log files, so care must be taken in dealing with raw
    logs.</p>
  </section>
  
  <section id="errorlog">
    <title>Error Log</title>
    
    <related>
      <directivelist>
        <directive module="mod_core">ErrorLog</directive>
        <directive module="mod_core">LogLevel</directive>
      </directivelist>
    </related>
    
    <p>The server error log, whose name and location is set by the
    <directive module="core">ErrorLog</directive> directive, is the
    most important log file. This is the place where Apache httpd
    will send diagnostic information and record any errors that it
    encounters in processing requests. It is the first place to
    look when a problem occurs with starting the server or with the
    operation of the server, since it will often contain details of
    what went wrong and how to fix it.</p>

    <p>The error log is usually written to a file (typically
    <code>error_log</code> on unix systems and
    <code>error.log</code> on Windows and OS/2). On unix systems it
    is also possible to have the server send errors to
    <code>syslog</code> or <a href="#piped">pipe them to a
    program</a>.</p>
    
    <p>The format of the error log is relatively free-form and
    descriptive. But there is certain information that is contained
    in most error log entries. For example, here is a typical
    message.</p>
    
    <example>
      [Wed Oct 11 14:32:52 2000] [error] [client 127.0.0.1]
      client denied by server configuration:
      /export/home/live/ap/htdocs/test
    </example>
    
    <p>The first item in the log entry is the date and time of the
    message. The second entry lists the severity of the error being
    reported. The <directive module="core">LogLevel</directive>
    directive is used to control the types of errors that are sent
    to the error log by restricting the severity level. The third
    entry gives the IP address of the client that generated the
    error. Beyond that is the message itself, which in this case
    indicates that the server has been configured to deny the
    client access. The server reports the file-system path (as
    opposed to the web path) of the requested document.</p>

    <p>A very wide variety of different messages can appear in the
    error log. Most look similar to the example above. The error
    log will also contain debugging output from CGI scripts. Any
    information written to <code>stderr</code> by a CGI script will
    be copied directly to the error log.</p>

    <p>It is not possible to customize the error log by adding or
    removing information. However, error log entries dealing with
    particular requests have corresponding entries in the <a
    href="#accesslog">access log</a>. For example, the above example
    entry corresponds to an access log entry with status code 403.
    Since it is possible to customize the access log, you can
    obtain more information about error conditions using that log
    file.</p>
    
    <p>During testing, it is often useful to continuously monitor
    the error log for any problems. On unix systems, you can
    accomplish this using:</p>
    
    <example>
      tail -f error_log
    </example>
  </section>
  
  <section id="accesslog">
    <title>Access Log</title>
    
    <related>
      <modulelist>
        <module>mod_log_config</module>
        <module>mod_setenvif</module>
      </modulelist>
      <directivelist>
        <directive module="mod_log_config">CustomLog</directive>
        <directive module="mod_log_config">LogFormat</directive>
        <directive module="mod_setenvif">SetEnvIf</directive>
      </directivelist>
    </related>
    
    <p>The server access log records all requests processed by the
    server. The location and content of the access log are
    controlled by the <directive module="mod_log_config">CustomLog</directive>
    directive. The <directive module="mod_log_config">LogFormat</directive>
    directive can be used to simplify the selection of 
    the contents of the logs. This section describes how to configure the server
    to record information in the access log.</p>

    <p>Of course, storing the information in the access log is only
    the start of log management. The next step is to analyze this
    information to produce useful statistics. Log analysis in
    general is beyond the scope of this document, and not really
    part of the job of the web server itself. For more information
    about this topic, and for applications which perform log
    analysis, check the <a
    href="http://dmoz.org/Computers/Software/Internet/Site_Management/Log_analysis/">
    Open Directory</a> or <a
    href="http://dir.yahoo.com/Computers_and_Internet/Software/Internet/World_Wide_Web/Servers/Log_Analysis_Tools/">
    Yahoo</a>.</p>
    
    <p>Various versions of Apache httpd have used other modules and
    directives to control access logging, including
    mod_log_referer, mod_log_agent, and the
    <code>TransferLog</code> directive. The <code>CustomLog</code>
    directive now subsumes the functionality of all the older
    directives.</p>

    <p>The format of the access log is highly configurable. The
    format is specified using a <directive module="mod_log_config">
    CustomLog</directive> that
    looks much like a C-style printf(1) format string. Some
    examples are presented in the next sections. For a complete
    list of the possible contents of the format string, see the <a
    href="mod/mod_log_config.html">mod_log_config
    documentation</a>.</p>
    
    <section id="common">
      <title>Common Log Format</title>
    
      <p>A typical configuration for the access log might look as
      follows.</p>

      <example>
        LogFormat "%h %l %u %t \"%r\" %&gt;s %b" common<br />
         CustomLog logs/access_log common
      </example>
    
      <p>This defines the <em>nickname</em> <code>common</code> and
      associates it with a particular log format string. The format
      string consists of percent directives, each of which tell the
      server to log a particular piece of information. Literal
      characters may also be placed in the format string and will be
      copied directly into the log output. The quote character
      (<code>"</code>) must be escaped by placing a back-slash before
      it to prevent it from being interpreted as the end of the
      format string. The format string may also contain the special
      control characters "<code>\n</code>" for new-line and
      "<code>\t</code>" for tab.</p>

      <p>The <code>CustomLog</code> directive sets up a new log file
      using the defined <em>nickname</em>. The filename for the
      access log is relative to the <directive
      module="core">ServerRoot</directive> unless it begins
      with a slash.</p>
    
      <p>The above configuration will write log entries in a format
      known as the Common Log Format (CLF). This standard format can
      be produced by many different web servers and read by many log
      analysis programs. The log file entries produced in CLF will
      look something like this:</p>

      <example>
        127.0.0.1 - frank [10/Oct/2000:13:55:36 -0700] "GET
        /apache_pb.gif HTTP/1.0" 200 2326
      </example>
    
      <p>Each part of this log entry is described below.</p>
    
      <dl>
        <dt><code>127.0.0.1</code> (<code>%h</code>)</dt>

        <dd>This is the IP address of the client (remote host) which
        made the request to the server. If <directive
        module="core">HostnameLookups</directive> is
        set to <code>On</code>, then the server will try to determine
        the hostname and log it in place of the IP address. However,
        this configuration is not recommended since it can
        significantly slow the server. Instead, it is best to use a
        log post-processor such as <a
        href="programs/logresolve.html">logresolve</a> to determine
        the hostnames. The IP address reported here is not
        necessarily the address of the machine at which the user is
        sitting. If a proxy server exists between the user and the
        server, this address will be the address of the proxy, rather
        than the originating machine.</dd>

        <dt><code>-</code> (<code>%l</code>)</dt>

        <dd>The "hyphen" in the output indicates that the requested
        piece of information is not available. In this case, the
        information that is not available is the RFC 1413 identity of
        the client determined by <code>identd</code> on the clients
        machine. This information is highly unreliable and should
        almost never be used except on tightly controlled internal
        networks. Apache httpd will not even attempt to determine
        this information unless <directive
        module="core">IdentityCheck</directive> is set
        to <code>On</code>.</dd>

        <dt><code>frank</code> (<code>%u</code>)</dt>

        <dd>This is the userid of the person requesting the document
        as determined by HTTP authentication. The same value is
        typically provided to CGI scripts in the
        <code>REMOTE_USER</code> environment variable. If the status
        code for the request (see below) is 401, then this value
        should not be trusted because the user is not yet
        authenticated. If the document is not password protected,
        this entry will be "<code>-</code>" just like the previous
        one.</dd>

        <dt><code>[10/Oct/2000:13:55:36 -0700]</code>
        (<code>%t</code>)</dt>

        <dd>
          The time that the server finished processing the request.
          The format is: 

          <blockquote>
            <code>[day/month/year:hour:minute:second zone]<br />
             day = 2*digit<br />
             month = 3*letter<br />
             year = 4*digit<br />
             hour = 2*digit<br />
             minute = 2*digit<br />
             second = 2*digit<br />
             zone = (`+' | `-') 4*digit</code>
          </blockquote>
          It is possible to have the time displayed in another format
          by specifying <code>%{format}t</code> in the log format
          string, where <code>format</code> is as in
          <code>strftime(3)</code> from the C standard library.
        </dd>

        <dt><code>"GET /apache_pb.gif HTTP/1.0"</code>
        (<code>\"%r\"</code>)</dt>

        <dd>The request line from the client is given in double
        quotes. The request line contains a great deal of useful
        information. First, the method used by the client is
        <code>GET</code>. Second, the client requested the resource
        <code>/apache_pb.gif</code>, and third, the client used the
        protocol <code>HTTP/1.0</code>. It is also possible to log
        one or more parts of the request line independently. For
        example, the format string "<code>%m %U%q %H</code>" will log
        the method, path, query-string, and protocol, resulting in
        exactly the same output as "<code>%r</code>".</dd>

        <dt><code>200</code> (<code>%&gt;s</code>)</dt>

        <dd>This is the status code that the server sends back to the
        client. This information is very valuable, because it reveals
        whether the request resulted in a successful response (codes
        beginning in 2), a redirection (codes beginning in 3), an
        error caused by the client (codes beginning in 4), or an
        error in the server (codes beginning in 5). The full list of
        possible status codes can be found in the <a
        href="http://www.w3.org/Protocols/rfc2616/rfc2616.txt">HTTP
        specification</a> (RFC2616 section 10).</dd>

        <dt><code>2326</code> (<code>%b</code>)</dt>

        <dd>The last entry indicates the size of the object returned
        to the client, not including the response headers. If no
        content was returned to the client, this value will be
        "<code>-</code>". To log "<code>0</code>" for no content, use
        <code>%B</code> instead.</dd>
      </dl>
    </section>
    
    <section id="combined">
      <title>Combined Log Format</title>
      
      <p>Another commonly used format string is called the Combined
      Log Format. It can be used as follows.</p>

      <example>
        LogFormat "%h %l %u %t \"%r\" %&gt;s %b \"%{Referer}i\"
        \"%{User-agent}i\"" combined<br />
         CustomLog log/acces_log combined
      </example>

      <p>This format is exactly the same as the Common Log Format,
      with the addition of two more fields. Each of the additional
      fields uses the percent-directive
      <code>%{<em>header</em>}i</code>, where <em>header</em> can be
      any HTTP request header. The access log under this format will
      look like:</p>
      
      <example>
        127.0.0.1 - frank [10/Oct/2000:13:55:36 -0700] "GET
        /apache_pb.gif HTTP/1.0" 200 2326
        "http://www.example.com/start.html" "Mozilla/4.08 [en]
        (Win98; I ;Nav)"
      </example>

      <p>The additional fields are:</p>
      
      <dl>
        <dt><code>"http://www.example.com/start.html"</code>
        (<code>\"%{Referer}i\"</code>)</dt>

        <dd>The "Referer" (sic) HTTP request header. This gives the
        site that the client reports having been referred from. (This
        should be the page that links to or includes
        <code>/apache_pb.gif</code>).</dd>

        <dt><code>"Mozilla/4.08 [en] (Win98; I ;Nav)"</code>
        (<code>\"%{User-agent}i\"</code>)</dt>

        <dd>The User-Agent HTTP request header. This is the
        identifying information that the client browser reports about
        itself.</dd>
      </dl>
    </section>
    
    <section id="multiple">
      <title>Multiple Access Logs</title>
      
      <p>Multiple access logs can be created simply by specifying
      multiple <directive module="core">CustomLog</directive> 
      directives in the configuration
      file. For example, the following directives will create three
      access logs. The first contains the basic CLF information,
      while the second and third contain referer and browser
      information. The last two <directive
      module="core">CustomLog</directive> lines show how
      to mimic the effects of the <directive module="core">ReferLog</directive> 
      and <directive module="core">AgentLog</directive> directives.</p>

      <example>
        LogFormat "%h %l %u %t \"%r\" %&gt;s %b" common<br />
        CustomLog logs/access_log common<br />
        CustomLog logs/referer_log "%{Referer}i -&gt; %U"<br />
        CustomLog logs/agent_log "%{User-agent}i"
      </example>

      <p>This example also shows that it is not necessary to define a
      nickname with the <code>LogFormat</code> directive. Instead,
      the log format can be specified directly in the 
      <directive module="core">CustomLog</directive> directive.</p>
    </section>
    
    <section id="conditional">
      <title>Conditional Logs</title>
      
      <p>There are times when it is convenient to exclude certain
      entries from the access logs based on characteristics of the
      client request. This is easily accomplished with the help of <a
      href="env.html">environment variables</a>. First, an
      environment variable must be set to indicate that the request
      meets certain conditions. This is usually accomplished with 
      <directive module="mod_setenvif">SetEnvIf</directive>. Then the
      <code>env=</code> clause of the <code>CustomLog</code>
      directive is used to include or exclude requests where the
      environment variable is set. Some examples:</p>

      <example>
        # Mark requests from the loop-back interface<br />
        SetEnvIf Remote_Addr "127\.0\.0\.1" dontlog<br />
        # Mark requests for the robots.txt file<br />
        SetEnvIf Request_URI "^/robots\.txt$" dontlog<br />
        # Log what remains<br />
        CustomLog logs/access_log common env=!dontlog
      </example>

      <p>As another example, consider logging requests from
      english-speakers to one log file, and non-english speakers to a
      different log file.</p>

      <example>
        SetEnvIf Accept-Language "en" english<br />
        CustomLog logs/english_log common env=english<br />
        CustomLog logs/non_english_log common env=!english
      </example>

      <p>Although we have just shown that conditional logging is very
      powerful and flexibly, it is not the only way to control the
      contents of the logs. Log files are more useful when they
      contain a complete record of server activity. It is often
      easier to simply post-process the log files to remove requests
      that you do not want to consider.</p>
    </section>
  </section>
  
  <section id="rotation">
    <title>Rotation Logs</title>
    
    <p>On even a moderately busy server, the quantity of
    information stored in the log files is very large. The access
    log file typically grows 1 MB or more per 10,000 requests. It
    will consequently be necessary to periodically rotate the log
    files by moving or deleting the existing logs. This cannot be
    done while the server is running, because Apache will continue
    writing to the old log file as long as it holds the file open.
    Instead, the server must be <a
    href="stopping.html">restarted</a> after the log files are
    moved or deleted so that it will open new log files.</p>

    <p>By using a <em>graceful</em> restart, the server can be
    instructed to open new log files without losing any existing or
    pending connections from clients. However, in order to
    accomplish this, the server must continue to write to the old
    log files while it finishes serving old requests. It is
    therefore necessary to wait for some time after the restart
    before doing any processing on the log files. A typical
    scenario that simply rotates the logs and compresses the old
    logs to save space is:</p>

    <example>
      mv access_log access_log.old<br />
      mv error_log error_log.old<br />
      apachectl graceful<br />
      sleep 600<br />
      gzip access_log.old error_log.old
    </example>

    <p>Another way to perform log rotation is using <a
    href="#piped">piped logs</a> as discussed in the next
    section.</p>
  </section>
  
  <section id="piped">
    <title>Piped Logs</title>
    
    <p>Apache httpd is capable of writing error and access log
    files through a pipe to another process, rather than directly
    to a file. This capability dramatically increases the
    flexibility of logging, without adding code to the main server.
    In order to write logs to a pipe, simply replace the filename
    with the pipe character "<code>|</code>", followed by the name
    of the executable which should accept log entries on its
    standard input. Apache will start the piped-log process when
    the server starts, and will restart it if it crashes while the
    server is running. (This last feature is why we can refer to
    this technique as "reliable piped logging".)</p>

    <p>Piped log processes are spawned by the parent Apache httpd
    process, and inherit the userid of that process. This means
    that piped log programs usually run as root. It is therefore
    very important to keep the programs simple and secure.</p>

    <p>Some simple examples using piped logs:</p>

    <example>
      # compressed logs<br />
      CustomLog "|/usr/bin/gzip -c &gt;&gt;
      /var/log/access_log.gz" common<br />
      # almost-real-time name resolution<br />
      CustomLog "|/usr/local/apache/bin/logresolve &gt;&gt;
      /var/log/access_log" common
    </example>

    <p>Notice that quotes are used to enclose the entire command
    that will be called for the pipe. Although these examples are
    for the access log, the same technique can be used for the
    error log.</p>

    <p>One important use of piped logs is to allow log rotation
    without having to restart the server. The Apache HTTP Server
    includes a simple program called <a
    href="programs/rotatelogs.html">rotatelogs</a> for this
    purpose. For example, to rotate the logs every 24 hours, you
    can use:</p>

    <example>
      CustomLog "|/usr/local/apache/bin/rotatelogs
      /var/log/access_log 86400" common
    </example>
    
    <p>A similar, but much more flexible log rotation program
    called <a href="http://www.cronolog.org/">cronolog</a>
    is available at an external site.</p>

    <p>As with conditional logging, piped logs are a very powerful
    tool, but they should not be used where a simpler solution like
    off-line post-processing is available.</p>
  </section>
  
  <section id="virtualhost">
    <title>Virtual Hosts</title>
    
    <p>When running a server with many <a href="vhosts/">virtual
    hosts</a>, there are several options for dealing with log
    files. First, it is possible to use logs exactly as in a
    single-host server. Simply by placing the logging directives
    outside the <directive module="core" 
    type="section">VirtualHost</directive> sections in the
    main server context, it is possible to log all requests in the
    same access log and error log. This technique does not allow
    for easy collection of statistics on individual virtual
    hosts.</p>

    <p>If <directive module="mod_log_config">CustomLog</directive> 
    or <directive module="mod_log_config">ErrorLog</directive>
    directives are placed inside a
    <directive module="core" type="section">VirtualHost</directive>
    section, all requests or errors for that virtual host will be
    logged only to the specified file. Any virtual host which does
    not have logging directives will still have its requests sent
    to the main server logs. This technique is very useful for a
    small number of virtual hosts, but if the number of hosts is
    very large, it can be complicated to manage. In addition, it
    can often create problems with <a
    href="vhosts/fd-limits.html">insufficient file
    descriptors</a>.</p>

    <p>For the access log, there is a very good compromise. By
    adding information on the virtual host to the log format
    string, it is possible to log all hosts to the same log, and
    later split the log into individual files. For example,
    consider the following directives.</p>

    <example>
      LogFormat "%v %l %u %t \"%r\" %&gt;s %b"
      comonvhost<br />
      CustomLog logs/access_log comonvhost
    </example>

    <p>The <code>%v</code> is used to log the name of the virtual
    host that is serving the request. Then a program like <a
    href="programs/other.html">split-logfile</a> can be used to
    post-process the access log in order to split it into one file
    per virtual host.</p>
  </section>
  
  <section id="other">
    <title>Other Log Files</title>
    
    <related>
      <modulelist>
        <module>mod_cgi</module>
        <module>mod_rewrite</module>
      </modulelist>
      <directivelist>
        <directive module="core">PidFile</directive>
        <directive module="mod_rewrite">RewriteLog</directive>
        <directive module="mod_rewrite">RewriteLogLevel</directive>
        <directive module="mod_cgi">ScriptLog</directive>
        <directive module="mod_cgi">ScriptLogLength</directive>
        <directive module="mod_cgi">ScriptBuffer</directive>
      </directivelist>
    </related>
    
    <section id="pidfile">
      <title>PID File</title>
    
      <p>On startup, Apache httpd saves the process id of the parent
      httpd process to the file <code>logs/httpd.pid</code>. This
      filename can be changed with the <directive module="core">PidFile
      </directive> directive. The
      process-id is for use by the administrator in restarting and
      terminating the daemon by sending signals to the parent
      process; on Windows, use the -k command line option instead.
      For more information see the <a href="stopping.html">Stopping
      and Restarting</a> page.</p>    
    </section>
    
    <section id="scriptlog">
      <title>Script Log</title>
      
      <p>In order to aid in debugging, the
      <directive module="mod_cgi">ScriptLog</directive> directive
      allows you to record the input to and output from CGI scripts.
      This should only be used in testing - not for live servers.
      More information is available in the <a
      href="mod/mod_cgi.html">mod_cgi documentation</a>.</p>
    </section>
    
    <section id="rewritelog">
      <title>Rewrite Log</title>
      
      <p>When using the powerful and complex features of <a
      href="mod/mod_rewrite.html">mod_rewrite</a>, it is almost
      always necessary to use the <directive
      module="mod_rewrite">RewriteLog</directive> to help
      in debugging. This log file produces a detailed analysis of how
      the rewriting engine transforms requests. The level of detail
      is controlled by the <directive
      module="mod_rewrite">RewriteLogLevel</directive> directive.</p>
    </section>
  </section>
</manualpage>