mass.xml revision 264920cbb38b10e7f60b75d0d1b894fa10bcbabc
<?xml version='1.0' encoding='UTF-8' ?>
<manualpage metafile="mass.xml.meta">
<parentdocument href="./">Virtual Hosts</parentdocument>
<title>Dynamically configured mass virtual hosting</title>
<summary>
<p>This document describes how to efficiently serve an
arbitrary number of virtual hosts with Apache 1.3. <!--
Written by Tony Finch (fanf@demon.net) (dot@dotat.at).
Some examples were derived from Ralf S. Engleschall's document
Some suggestions were made by Brian Behlendorf.
-->
</p>
</summary>
<section id="motivation"><title>Motivation</title>
<p>The techniques described here are of interest if your
<code><VirtualHost></code> sections that are
substantially the same, for example:</p>
<example>
NameVirtualHost 111.22.33.44<br />
<VirtualHost 111.22.33.44><br />
<indent>
ServerName www.customer-1.com<br />
</indent>
</VirtualHost><br />
<VirtualHost 111.22.33.44><br />
<indent>
ServerName www.customer-2.com<br />
</indent>
</VirtualHost><br />
# blah blah blah<br />
<VirtualHost 111.22.33.44><br />
<indent>
ServerName www.customer-N.com<br />
</indent>
</VirtualHost>
</example>
<p>The basic idea is to replace all of the static
<code><VirtualHost></code> configuration with a mechanism
that works it out dynamically. This has a number of
advantages:</p>
<ol>
<li>Your configuration file is smaller so Apache starts
faster and uses less memory.</li>
<li>Adding virtual hosts is simply a matter of creating the
appropriate directories in the filesystem and entries in the
DNS - you don't need to reconfigure or restart Apache.</li>
</ol>
<p>The main disadvantage is that you cannot have a different
log file for each virtual host; however if you have very many
virtual hosts then doing this is dubious anyway because it eats
file descriptors. It is better to log to a pipe or a fifo and
arrange for the process at the other end to distribute the logs
to the customers (it can also accumulate statistics, etc.).</p>
</section>
<section id="overview"><title>Overview</title>
<p>A virtual host is defined by two pieces of information: its
IP address, and the contents of the <code>Host:</code> header
in the HTTP request. The dynamic mass virtual hosting technique
is based on automatically inserting this information into the
pathname of the file that is used to satisfy the request. This
is done most easily using <module>mod_vhost_alias</module>,
but if you are using a version of Apache up to 1.3.6 then you
must use <module>mod_rewrite</module>.
Both of these modules are disabled by default; you must enable
one of them when configuring and building Apache if you want to
use this technique.</p>
<p>A couple of things need to be `faked' to make the dynamic
virtual host look like a normal one. The most important is the
server name which is used by Apache to generate
self-referential URLs, etc. It is configured with the
<code>ServerName</code> directive, and it is available to CGIs
via the <code>SERVER_NAME</code> environment variable. The
actual value used at run time is controlled by the <directive
module="core">UseCanonicalName</directive>
setting. With <code>UseCanonicalName Off</code> the server name
comes from the contents of the <code>Host:</code> header in the
request. With <code>UseCanonicalName DNS</code> it comes from a
reverse DNS lookup of the virtual host's IP address. The former
setting is used for name-based dynamic virtual hosting, and the
latter is used for IP-based hosting. If Apache cannot work out
the server name because there is no <code>Host:</code> header
or the DNS lookup fails then the value configured with
<code>ServerName</code> is used instead.</p>
<p>The other thing to `fake' is the document root (configured
with <code>DocumentRoot</code> and available to CGIs via the
<code>DOCUMENT_ROOT</code> environment variable). In a normal
configuration this setting is used by the core module when
mapping URIs to filenames, but when the server is configured to
do dynamic virtual hosting that job is taken over by another
module (either <code>mod_vhost_alias</code> or
<code>mod_rewrite</code>) which has a different way of doing
the mapping. Neither of these modules is responsible for
setting the <code>DOCUMENT_ROOT</code> environment variable so
if any CGIs or SSI documents make use of it they will get a
misleading value.</p>
</section>
<section id="simple"><title>Simple dynamic virtual hosts</title>
virtual host arrangement outlined in the <a
href="#motivation">Motivation</a> section above, but in a
generic fashion using <code>mod_vhost_alias</code>.</p>
<example>
# get the server name from the Host: header<br />
UseCanonicalName Off<br />
<br />
# this log format can be split per-virtual-host based on the first field<br />
LogFormat "%V %h %l %u %t \"%r\" %s %b" vcommon<br />
<br />
# include the server name in the filenames used to satisfy requests<br />
</example>
<p>This configuration can be changed into an IP-based virtual
hosting solution by just turning <code>UseCanonicalName
Off</code> into <code>UseCanonicalName DNS</code>. The server
name that is inserted into the filename is then derived from
the IP address of the virtual host.</p>
</section>
<section id="homepages"><title>A virtually hosted homepages system</title>
<p>This is an adjustment of the above system tailored for an
ISP's homepages server. Using a slightly more complicated
configuration we can select substrings of the server name to
use in the filename so that e.g. the documents for
directory instead of one per virtual host.</p>
<example>
# all the preliminary stuff is the same as above, then<br />
<br />
# include part of the server name in the filenames<br />
<br />
# single cgi-bin directory<br />
</example>
<p>There are examples of more complicated
<code>VirtualDocumentRoot</code> settings in the
<module>mod_vhost_alias</module> documentation.</p>
</section>
<section id="combinations"><title>Using more than
one virtual hosting system on the same server</title>
<p>With more complicated setups you can use Apache's normal
<code><VirtualHost></code> directives to control the
scope of the various virtual hosting configurations. For
example, you could have one IP address for homepages customers
and another for commercial customers with the following setup.
This can of course be combined with conventional
<code><VirtualHost></code> configuration sections.</p>
<example>
UseCanonicalName Off<br />
<br />
LogFormat "%V %h %l %u %t \"%r\" %s %b" vcommon<br />
<br />
<indent>
Options FollowSymLinks<br />
AllowOverride All<br />
</indent>
</Directory><br />
<br />
<indent>
Options FollowSymLinks<br />
AllowOverride None<br />
</indent>
</Directory><br />
<br />
<VirtualHost 111.22.33.44><br />
<indent>
ServerName www.commercial.isp.com<br />
<br />
<br />
</indent>
</VirtualHost><br />
<br />
<VirtualHost 111.22.33.45><br />
<indent>
ServerName www.homepages.isp.com<br />
<br />
<br />
</indent>
</VirtualHost>
</example>
</section>
<section id="ipbased"><title>More efficient IP-based virtual hosting</title>
<p>After <a href="#simple">the first example</a> I noted that
it is easy to turn it into an IP-based virtual hosting setup.
Unfortunately that configuration is not very efficient because
it requires a DNS lookup for every request. This can be avoided
by laying out the filesystem according to the IP addresses
themselves rather than the corresponding names and changing the
logging similarly. Apache will then usually not need to work
out the server name and so incur a DNS lookup.</p>
<example>
# get the server name from the reverse DNS of the IP address<br />
UseCanonicalName DNS<br />
<br />
# include the IP address in the logs so they may be split<br />
LogFormat "%A %h %l %u %t \"%r\" %s %b" vcommon<br />
<br />
# include the IP address in the filenames<br />
</example>
</section>
<section id="oldversion"><title>Using older versions of Apache</title>
<p>The examples above rely on <code>mod_vhost_alias</code>
which appeared after version 1.3.6. If you are using a version
of Apache without <code>mod_vhost_alias</code> then you can
implement this technique with <code>mod_rewrite</code> as
illustrated below, but only for Host:-header-based virtual
hosts.</p>
<p>In addition there are some things to beware of with logging.
Apache 1.3.6 is the first version to include the
<code>%V</code> log format directive; in versions 1.3.0 - 1.3.3
the <code>%v</code> option did what <code>%V</code> does;
version 1.3.4 has no equivalent. In all these versions of
Apache the <code>UseCanonicalName</code> directive can appear
in <code>.htaccess</code> files which means that customers can
cause the wrong thing to be logged. Therefore the best thing to
do is use the <code>%{Host}i</code> directive which logs the
<code>Host:</code> header directly; note that this may include
<code>:port</code> on the end which is not the case for
<code>%V</code>.</p>
</section>
virtual hosts using <code>mod_rewrite</code></title>
thing as <a href="#simple">the first example</a>. The first
half is very similar to the corresponding part above but with
some changes for backward compatibility and to make the
<code>mod_rewrite</code> part work properly; the second half
configures <code>mod_rewrite</code> to do the actual work.</p>
<p>There are a couple of especially tricky bits: By default,
<code>mod_rewrite</code> runs before the other URI translation
modules (<code>mod_alias</code> etc.) so if they are used then
<code>mod_rewrite</code> must be configured to accommodate
them. Also, some magic must be performed to do a
per-dynamic-virtual-host equivalent of
<code>ScriptAlias</code>.</p>
<example>
# get the server name from the Host: header<br />
UseCanonicalName Off<br />
<br />
# splittable logs<br />
LogFormat "%{Host}i %h %l %u %t \"%r\" %s %b" vcommon<br />
<br />
<indent>
# ExecCGI is needed here because we can't force<br />
# CGI execution in the way that ScriptAlias does<br />
Options FollowSymLinks ExecCGI<br />
</indent>
</Directory><br />
<br />
# now for the hard bit<br />
<br />
RewriteEngine On<br />
<br />
# a ServerName derived from a Host: header may be any case at all<br />
RewriteMap lowercase int:tolower<br />
<br />
## deal with normal documents first:<br />
# allow Alias /icons/ to work - repeat for other aliases<br />
RewriteCond %{REQUEST_URI} !^/icons/<br />
# allow CGIs to work<br />
RewriteCond %{REQUEST_URI} !^/cgi-bin/<br />
# do the magic<br />
<br />
## and now deal with CGIs - we have to force a MIME type<br />
RewriteCond %{REQUEST_URI} ^/cgi-bin/<br />
RewriteRule ^/(.*)$ /www/hosts/${lowercase:%{SERVER_NAME}}/cgi-bin/$1 [T=application/x-httpd-cgi]<br />
<br />
# that's it!
</example>
</section>
homepages system using <code>mod_rewrite</code></title>
<p>This does the same thing as <a href="#homepages">the second
example</a>.</p>
<example>
RewriteEngine on<br />
<br />
RewriteMap lowercase int:tolower<br />
<br />
# allow CGIs to work<br />
RewriteCond %{REQUEST_URI} !^/cgi-bin/<br />
<br />
# check the hostname is right so that the RewriteRule works<br />
RewriteCond ${lowercase:%{SERVER_NAME}} ^www\.[a-z-]+\.isp\.com$<br />
<br />
# concatenate the virtual host name onto the start of the URI<br />
# the [C] means do the next rewrite on the result of this one<br />
RewriteRule ^(.+) ${lowercase:%{SERVER_NAME}}$1 [C]<br />
<br />
# now create the real file name<br />
RewriteRule ^www\.([a-z-]+)\.isp\.com/(.*) /home/$1/$2<br />
<br />
# define the global CGI directory<br />
</example>
</section>
<section id="xtra-conf"><title>Using a separate virtual
host configuration file</title>
<p>This arrangement uses more advanced <code>mod_rewrite</code>
features to get the translation from virtual host to document
root from a separate configuration file. This provides more
flexibility but requires more complicated configuration.</p>
this:</p>
<example>
# ...<br />
</example>
<example>
RewriteEngine on<br />
<br />
RewriteMap lowercase int:tolower<br />
<br />
# define the map file<br />
<br />
# deal with aliases as above<br />
RewriteCond %{REQUEST_URI} !^/icons/<br />
RewriteCond %{REQUEST_URI} !^/cgi-bin/<br />
RewriteCond ${lowercase:%{SERVER_NAME}} ^(.+)$<br />
# this does the file-based remap<br />
RewriteCond ${vhost:%1} ^(/.*)$<br />
<br />
RewriteCond %{REQUEST_URI} ^/cgi-bin/<br />
RewriteCond ${lowercase:%{SERVER_NAME}} ^(.+)$<br />
RewriteCond ${vhost:%1} ^(/.*)$<br />
</example>
</section>
</manualpage>