<?xml version='1.0' encoding='UTF-8' ?>

<!DOCTYPE manualpage SYSTEM "/style/manualpage.dtd">

<?xml-stylesheet type="text/xsl" href="/style/manual.en.xsl"?>

<manualpage metafile="details.xml.meta">

<parentdocument href="./">Virtual Hosts</parentdocument>

<title>An In-Depth Discussion of Virtual Host Matching</title>

<summary>

<p>The virtual host code was completely rewritten in

<strong>Apache 1.3</strong>. This document attempts to explain

exactly what Apache does when deciding what virtual host to

serve a hit from. With the help of the new

<directive module="core">NameVirtualHost</directive>

directive virtual host configuration should be a lot easier and

safer than with versions prior to 1.3.</p>

<p>If you just want to <cite>make it work</cite> without

understanding how, here are <a href="examples.html">some

examples</a>.</p>

</summary>

<section id="configparsing"><title>Config File Parsing</title>

<p>There is a <em>main_server</em> which consists of all the

definitions appearing outside of

<code>&lt;VirtualHost&gt;</code> sections. There are virtual

servers, called <em>vhosts</em>, which are defined by

<directive type="section" module="core">VirtualHost</directive>

sections.</p>

<p>The directives

<directive module="core">ServerName</directive> and

<directive module="core">ServerPath</directive>

can appear anywhere within the definition of a server. However,

each appearance overrides the previous appearance (within that

server).</p>

<p>The main_server has no default

<code>ServerPath</code>, or <code>ServerAlias</code>. The

default <code>ServerName</code> is deduced from the server's IP

address.</p>

<p>Port numbers specified in the <code>VirtualHost</code> directive do

not influence what port numbers Apache will listen on, they only discriminate between

which <code>VirtualHost</code> will be selected to handle a request.</p>

<p>Each address appearing in the <code>VirtualHost</code>

directive can have an optional port. If the port is unspecified

it is treated as a wildcard port. The special port <code>*</code>

indicates a wildcard that matches any port. Collectively the

entire set of addresses (including multiple <code>A</code>

record results from DNS lookups) are called the vhost's

<em>address set</em>.</p>

<p>Unless a <directive module="core">NameVirtualHost</directive>

directive is used for the exact IP address and port pair in the

<code>VirtualHost</code> directive, Apache selects the best match

only on the basis of the IP address (or wildcard) and port number.

If there are multiple identical best matches, the first <code>VirtualHost</code>

appearing in the configuration file will be selected.</p>

<p>If you want Apache to <em>further</em> discriminate on the basis of the

HTTP <code>Host</code> header supplied by the client, the

<code>NameVirtualHost</code> directive <em>must</em> appear

with the exact IP address (or wildcard) and port pair used in a correspnding

set of <code>VirtualHost</code> directives.</p>

<p>The name-based virtual host selection occurs only after the a single IP-based

virtual host has been selected, and only considers the set of virtual hosts

the carry an identical IP address and port pair.</p>

<p>Hostnames can be used in place of IP addresses in a virtual host definition,

but it is resolved at startup and is not recommended.</p>

<p>Multiple <code>NameVirtualHost</code> directives can be used

each with a set of <code>VirtualHost</code> directives but only

one <code>NameVirtualHost</code> directive should be used for

each specific IP:port pair.</p>

<p>The ordering of <code>NameVirtualHost</code> and

<code>VirtualHost</code> directives is not important which

makes the following two examples identical (only the order of

the <code>VirtualHost</code> directives for <em>one</em>

address set is important, see below):</p>

<table><tr>

<td><example>

NameVirtualHost 111.22.33.44<br />

&lt;VirtualHost 111.22.33.44&gt;<br />

# server A<br />

...<br />

&lt;/VirtualHost&gt;<br />

&lt;VirtualHost 111.22.33.44&gt;<br />

# server B<br />

...<br />

&lt;/VirtualHost&gt;<br />

<br />

NameVirtualHost 111.22.33.55<br />

&lt;VirtualHost 111.22.33.55&gt;<br />

# server C<br />

...<br />

&lt;/VirtualHost&gt;<br />

&lt;VirtualHost 111.22.33.55&gt;<br />

# server D<br />

...<br />

&lt;/VirtualHost&gt;

</example></td>

<td><example>

&lt;VirtualHost 111.22.33.44&gt;<br />

# server A<br />

&lt;/VirtualHost&gt;<br />

&lt;VirtualHost 111.22.33.55&gt;<br />

# server C<br />

...<br />

&lt;/VirtualHost&gt;<br />

&lt;VirtualHost 111.22.33.44&gt;<br />

# server B<br />

...<br />

&lt;/VirtualHost&gt;<br />

&lt;VirtualHost 111.22.33.55&gt;<br />

# server D<br />

...<br />

&lt;/VirtualHost&gt;<br />

<br />

NameVirtualHost 111.22.33.44<br />

NameVirtualHost 111.22.33.55<br />

<br />

</example></td>

</tr></table>

<p>(To aid the readability of your configuration you should

prefer the left variant.)</p>

<p>During initialization a list for each IP address is

generated and inserted into an hash table. If the IP address is

used in a <code>NameVirtualHost</code> directive the list

contains all name-based vhosts for the given IP address. If

there are no vhosts defined for that address the

<code>NameVirtualHost</code> directive is ignored and an error

is logged. For an IP-based vhost the list in the hash table is

empty.</p>

<p>Due to a fast hashing function the overhead of hashing an IP

address during a request is minimal and almost not existent.

Additionally the table is optimized for IP addresses which vary

in the last octet.</p>

<p>For every vhost various default values are set. In

particular:</p>

<ol>

<li>If a vhost has no <directive module="core">ServerAdmin</directive>,

<directive module="core">ResourceConfig</directive>,

<directive module="core">AccessConfig</directive>,

<directive module="core">Timeout</directive>,

<directive module="core">KeepAliveTimeout</directive>,

<directive module="core">KeepAlive</directive>,

<directive module="core">MaxKeepAliveRequests</directive>,

<directive module="core">ReceiveBufferSize</directive>,

or <directive module="core">SendBufferSize</directive>

directive then the respective value is inherited from the

main_server. (That is, inherited from whatever the final

setting of that value is in the main_server.)</li>

<li>The "lookup defaults" that define the default directory

permissions for a vhost are merged with those of the

main_server. This includes any per-directory configuration

information for any module.</li>

<li>The per-server configs for each module from the

main_server are merged into the vhost server.</li>

</ol>

<p>Essentially, the main_server is treated as "defaults" or a

"base" on which to build each vhost. But the positioning of

these main_server definitions in the config file is largely

irrelevant -- the entire config of the main_server has been

parsed when this final merging occurs. So even if a main_server

definition appears after a vhost definition it might affect the

vhost definition.</p>

<p>If the main_server has no <code>ServerName</code> at this

point, then the hostname of the machine that <program>httpd</program>

is running on is used instead. We will call the <em>main_server address

set</em> those IP addresses returned by a DNS lookup on the

<code>ServerName</code> of the main_server.</p>

<p>For any undefined <code>ServerName</code> fields, a

name-based vhost defaults to the address given first in the

<code>VirtualHost</code> statement defining the vhost.</p>

<p>Any vhost that includes the magic <code>_default_</code>

wildcard is given the same <code>ServerName</code> as the

main_server.</p>

</section>

<section id="hostmatching"><title>Virtual Host Matching</title>

<p>The server determines which vhost to use for a request as

follows:</p>

<section id="hashtable"><title>Hash table lookup</title>

<p>When the connection is first made by a client, the IP

address to which the client connected is looked up in the

internal IP hash table.</p>

<p>If the lookup fails (the IP address wasn't found) the

request is served from the <code>_default_</code> vhost if

there is such a vhost for the port to which the client sent the

request. If there is no matching <code>_default_</code> vhost

the request is served from the main_server.</p>

<p>If the IP address is not found in the hash table then the

match against the port number may also result in an entry

corresponding to a <code>NameVirtualHost *</code>, which is

subsequently handled like other name-based vhosts.</p>

<p>If the lookup succeeded (a corresponding list for the IP

address was found) the next step is to decide if we have to

deal with an IP-based or a name-base vhost.</p>

</section>

<section id="ipbased"><title>IP-based vhost</title>

<p>If the entry we found has an empty name list then we have

found an IP-based vhost, no further actions are performed and

the request is served from that vhost.</p>

</section>

<section id="namebased"><title>Name-based vhost</title>

<p>If the entry corresponds to a name-based vhost the name list

contains one or more vhost structures. This list contains the

vhosts in the same order as the <code>VirtualHost</code>

directives appear in the config file.</p>

<p>The first vhost on this list (the first vhost in the config

file with the specified IP address) has the highest priority

and catches any request to an unknown server name or a request

without a <code>Host:</code> header field.</p>

<p>If the client provided a <code>Host:</code> header field the

list is searched for a matching vhost and the first hit on a

<code>ServerName</code> or <code>ServerAlias</code> is taken

and the request is served from that vhost. A <code>Host:</code>

header field can contain a port number, but Apache always

matches against the real port to which the client sent the

request.</p>

<p>If the client submitted a HTTP/1.0 request without

<code>Host:</code> header field we don't know to what server

the client tried to connect and any existing

<code>ServerPath</code> is matched against the URI from the

request. The first matching path on the list is used and the

request is served from that vhost.</p>

<p>If no matching vhost could be found the request is served

from the first vhost with a matching port number that is on the

list for the IP to which the client connected (as already

mentioned before).</p>

</section>

<section id="persistent"><title>Persistent connections</title>

<p>The IP lookup described above is only done <em>once</em> for a

particular TCP/IP session while the name lookup is done on

<em>every</em> request during a KeepAlive/persistent

connection. In other words a client may request pages from

different name-based vhosts during a single persistent

connection.</p>

</section>

<section id="absoluteURI"><title>Absolute URI</title>

<p>If the URI from the request is an absolute URI, and its

hostname and port match the main server or one of the

configured virtual hosts <em>and</em> match the address and

port to which the client sent the request, then the

scheme/hostname/port prefix is stripped off and the remaining

relative URI is served by the corresponding main server or

virtual host. If it does not match, then the URI remains

untouched and the request is taken to be a proxy request.</p>

</section>

<section id="observations"><title>Observations</title>

<ul>

<li>A name-based vhost can never interfere with an IP-base

vhost and vice versa. IP-based vhosts can only be reached

through an IP address of its own address set and never

through any other address. The same applies to name-based

vhosts, they can only be reached through an IP address of the

corresponding address set which must be defined with a

<code>NameVirtualHost</code> directive.</li>

<li><code>ServerAlias</code> and <code>ServerPath</code>

checks are never performed for an IP-based vhost.</li>

<li>The order of name-/IP-based, the <code>_default_</code>

vhost and the <code>NameVirtualHost</code> directive within

the config file is not important. Only the ordering of

name-based vhosts for a specific address set is significant.

The one name-based vhosts that comes first in the

configuration file has the highest priority for its

corresponding address set.</li>

<li>The <code>Host:</code> header field is never used during the

matching process. Apache always uses the real port to which

the client sent the request.</li>

<li>If a <code>ServerPath</code> directive exists which is a

prefix of another <code>ServerPath</code> directive that

appears later in the configuration file, then the former will

always be matched and the latter will never be matched. (That

is assuming that no <code>Host:</code> header field was

available to disambiguate the two.)</li>

<li>If two IP-based vhosts have an address in common, the

vhost appearing first in the config file is always matched.

Such a thing might happen inadvertently. The server will give

a warning in the error logfile when it detects this.</li>

<li>A <code>_default_</code> vhost catches a request only if

there is no other vhost with a matching IP address

<em>and</em> a matching port number for the request. The

request is only caught if the port number to which the client

sent the request matches the port number of your

<code>_default_</code> vhost which is your standard

<code>Listen</code> by default. A wildcard port can be

specified (<em>i.e.</em>, <code>_default_:*</code>) to catch

requests to any available port. This also applies to

<code>NameVirtualHost *</code> vhosts. Note that this is simply an

extension of the "best match" principle, as a specific and exact match

is favored over a wildcard.</li>

<li>The main_server is only used to serve a request if the IP

address and port number to which the client connected is

unspecified and does not match any other vhost (including a

<code>_default_</code> vhost). In other words the main_server

only catches a request for an unspecified address/port

combination (unless there is a <code>_default_</code> vhost

which matches that port).</li>

<li>A <code>_default_</code> vhost or the main_server is

<em>never</em> matched for a request with an unknown or

missing <code>Host:</code> header field if the client

connected to an address (and port) which is used for

name-based vhosts, <em>e.g.</em>, in a

<code>NameVirtualHost</code> directive.</li>

<li>You should never specify DNS names in

<code>VirtualHost</code> directives because it will force

your server to rely on DNS to boot. Furthermore it poses a

security threat if you do not control the DNS for all the

domains listed. There's <a href="/dns-caveats.html">more

information</a> available on this and the next two

topics.</li>

<li><code>ServerName</code> should always be set for each

vhost. Otherwise A DNS lookup is required for each

vhost.</li>

</ul>

</section>

</section>

<section id="tips"><title>Tips</title>

<p>In addition to the tips on the <a

href="/dns-caveats.html#tips">DNS Issues</a> page, here are

some further tips:</p>

<ul>

<li>Place all main_server definitions before any

<code>VirtualHost</code> definitions. (This is to aid the

readability of the configuration -- the post-config merging

process makes it non-obvious that definitions mixed in around

virtual hosts might affect all virtual hosts.)</li>

<li>Group corresponding <code>NameVirtualHost</code> and

<code>VirtualHost</code> definitions in your configuration to

ensure better readability.</li>

<li>Avoid <code>ServerPaths</code> which are prefixes of

other <code>ServerPaths</code>. If you cannot avoid this then

you have to ensure that the longer (more specific) prefix

vhost appears earlier in the configuration file than the

shorter (less specific) prefix (<em>i.e.</em>, "ServerPath

/abc" should appear after "ServerPath /abc/def").</li>

</ul>

</section>

</manualpage>