Mercurial > hg > nginx-vendor-current

view html/ngx_core_module.html @ 640:eb208e0cf44d NGINX_1_1_4

Find changesets by keywords (author, files, the commit message), revision number or hash, or revset expression.
nginx 1.1.4 *) Feature: the ngx_http_upstream_keepalive module. *) Feature: the "proxy_http_version" directive. *) Feature: the "fastcgi_keep_conn" directive. *) Feature: the "worker_aio_requests" directive. *) Bugfix: if nginx was built --with-file-aio it could not be run on Linux kernel which did not support AIO. *) Bugfix: in Linux AIO error processing. Thanks to Hagai Avrahami. *) Bugfix: reduced memory consumption for long-lived requests. *) Bugfix: the module ngx_http_mp4_module did not support 64-bit MP4 "co64" atom.
author Igor Sysoev <http://sysoev.ru>
date Tue, 20 Sep 2011 00:00:00 +0400
parents f5a8cf31a203
children
line wrap: on
line source

<html><head><meta http-equiv="Content-Type" content="text/html; charset=utf-8"><title>Core Module</title></head><body><center><h3>Core Module</h3></center><a name="example"></a><center><h4>Example Configuration</h4></center><p><blockquote><pre>
user www www;
worker_processes 2;

error_log /var/log/nginx-error.log info;

events {
    use kqueue;
    worker_connections 2048;
}

...
</pre></blockquote></p><a name="directives"></a><center><h4>Directives</h4></center><hr><a name="daemon"></a><strong>syntax</strong>:
         <code>daemon <code>on</code> | <code>off</code></code><br><strong>default</strong>:
      <code>daemon on</code><br><strong>context</strong>:
      <code>main</code><br><p>
Determines whether nginx should become a daemon.
Mainly used during development.
</p><hr><a name="env"></a><strong>syntax</strong>:
         <code>env <code><i>VAR</i></code>[=<code><i>VALUE</i></code>]</code><br><strong>default</strong>:
      <code>env TZ</code><br><strong>context</strong>:
      <code>main</code><br><p>
Allows to limit a set of environment variables, change their values,
or create new environment variables, for the following cases:
<ul><li>
variable inheritance during a
<a href="control.html#upgrade">live upgrade</a>
of an executable file;
</li><li>
use of variables by the module
<a href="http/ngx_http_perl_module.html">ngx_http_perl_module</a>;
</li><li>
use of variables by worker processes.
Please bear in mind that controlling system libraries in this way
is not always possible as it is not uncommon for libraries to check
variables only during initialization, well before they can be set
using this directive.
An exception from this is an above mentioned
<a href="control.html#upgrade">live upgrade</a>
of an executable file.
</li></ul></p><p>
The TZ variable is always inherited and made available to the module
<a href="http/ngx_http_perl_module.html">ngx_http_perl_module</a>,
unless configured explicitly.
</p><p>
Usage example:
<blockquote><pre>
env MALLOC_OPTIONS;
env PERL5LIB=/data/site/modules;
env OPENSSL_ALLOW_PROXY_CERTS=1;
</pre></blockquote></p><hr><a name="include"></a><strong>syntax</strong>:
         <code>include <code><i>file</i></code> | <code><i>mask</i></code></code><br><strong>default</strong>:
      <strong>none</strong><br><strong>context</strong>:
      <strong>any</strong><br><p>
Includes another <code><i>file</i></code>, or files matching the
specified <code><i>mask</i></code>, into configuration.
Included files should consist of
syntactically correct directives and blocks.
</p><p>
Usage example:
<blockquote><pre>
include mime.types;
include vhosts/*.conf;
</pre></blockquote></p><hr><a name="master_process"></a><strong>syntax</strong>:
         <code>master_process <code>on</code> | <code>off</code></code><br><strong>default</strong>:
      <code>master_process on</code><br><strong>context</strong>:
      <code>main</code><br><p>
Determines whether worker processes are started.
This directive is intended for nginx developers.
</p><hr><a name="pid"></a><strong>syntax</strong>:
         <code>pid <code><i>file</i></code></code><br><strong>default</strong>:
      <code>pid nginx.pid</code><br><strong>context</strong>:
      <code>main</code><br><p>
Defines a <code><i>file</i></code> which will store the process ID of the main process.
</p><hr><a name="ssl_engine"></a><strong>syntax</strong>:
         <code>ssl_engine <code><i>device</i></code></code><br><strong>default</strong>:
      <strong>none</strong><br><strong>context</strong>:
      <code>main</code><br><p>
Defines the name of the hardware SSL accelerator.
</p><hr><a name="user"></a><strong>syntax</strong>:
         <code>user <code><i>user</i></code> [<code><i>group</i></code>]</code><br><strong>default</strong>:
      <code>user nobody nobody</code><br><strong>context</strong>:
      <code>main</code><br><p>
Defines <code><i>user</i></code> and <code><i>group</i></code>
credentials used by worker processes.
If <code><i>group</i></code> is omitted, a group whose name equals
that of <code><i>user</i></code> is used.
</p><hr><a name="timer_resolution"></a><strong>syntax</strong>:
         <code>timer_resolution <code><i>interval</i></code></code><br><strong>default</strong>:
      <strong>none</strong><br><strong>context</strong>:
      <code>main</code><br><p>
Reduces timer resolution in worker processes, thus reducing the
number of <code>gettimeofday()</code> system calls made.
By default, <code>gettimeofday()</code> is called each time
on receiving a kernel event.
With reduced resolution, <code>gettimeofday()</code> is only
called once per specified <code><i>interval</i></code>.
</p><p>
Example:
<blockquote><pre>
timer_resolution 100ms;
</pre></blockquote></p><p>
An internal implementation of interval depends on the method used:
<ul><li>
an <code>EVFILT_TIMER</code> filter if <code>kqueue</code> is used;
</li><li><code>timer_create()</code> if <code>eventport</code> is used;
</li><li><code>setitimer()</code> otherwise.
</li></ul></p><hr><a name="worker_rlimit_core"></a><strong>syntax</strong>:
         <code>worker_rlimit_core <code><i>size</i></code></code><br><strong>default</strong>:
      <strong>none</strong><br><strong>context</strong>:
      <code>main</code><br><p>
Changes the limit on the largest size of a core file
(<code>RLIMIT_CORE</code>) for worker processes.
Used to increase the limit without restarting the main process.
</p><hr><a name="worker_rlimit_nofile"></a><strong>syntax</strong>:
         <code>worker_rlimit_nofile <code><i>number</i></code></code><br><strong>default</strong>:
      <strong>none</strong><br><strong>context</strong>:
      <code>main</code><br><p>
Changes the limit on the maximum number of open files
(<code>RLIMIT_NOFILE</code>) for worker processes.
Used to increase the limit without restarting the main process.
</p><hr><a name="worker_priority"></a><strong>syntax</strong>:
         <code>worker_priority <code><i>number</i></code></code><br><strong>default</strong>:
      <code>worker_priority 0</code><br><strong>context</strong>:
      <code>main</code><br><p>
Defines a scheduling priority for worker processes like is
done by the <code>nice</code>: a negative
<code><i>number</i></code>
means higher priority.
Allowed range normally varies from -20 to 20.
</p><p>
Example:
<blockquote><pre>
worker_priority -10;
</pre></blockquote></p><hr><a name="worker_processes"></a><strong>syntax</strong>:
         <code>worker_processes <code><i>number</i></code></code><br><strong>default</strong>:
      <code>worker_processes 1</code><br><strong>context</strong>:
      <code>main</code><br><p>
Defines the number of worker processes.
</p><hr><a name="working_directory"></a><strong>syntax</strong>:
         <code>working_directory <code><i>directory</i></code></code><br><strong>default</strong>:
      <strong>none</strong><br><strong>context</strong>:
      <code>main</code><br><p>
Defines a current working directory for a worker process.
It is primarily used when writing a core-file, in which case
a working process should have write permission for the
specified directory.
</p></body></html>