Dynamically Configured Mass Virtual Hosting

Motivation

The techniques described here are of interest if your httpd.conf contains many <VirtualHost> sections that are substantially the same, for example:

The basic idea is to replace all of the static <VirtualHost> configurations with a mechanism that works them out dynamically. This has a number of advantages:

1. Your configuration file is smaller, so Apache starts more quickly and uses less memory.
2. 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.

The main disadvantage is that you cannot have a different log file for each virtual host; however, if you have many virtual hosts, doing this can be a bad idea anyway, because of the number of file descriptors needed. 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. (This can also be used to accumulate statistics, etc.).

Overview

A virtual host is defined by two pieces of information: its IP address, and the contents of the Host: header in the HTTP request. The dynamic mass virtual hosting technique used here is based on automatically inserting this information into the pathname of the file that is used to satisfy the request. This can be most easily done by using mod_vhost_alias with Apache 2.0. Alternatively, mod_rewrite can be used. 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.

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 ServerName directive, and it is available to CGIs via the SERVER_NAME environment variable. The actual value used at run time is controlled by the UseCanonicalName setting. With UseCanonicalName Off, the server name is taken from the contents of the Host: header in the request. With UseCanonicalName DNS, it is taken 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 Host: header, or the DNS lookup fails, then the value configured with ServerName is used instead.

The other thing to `fake' is the document root (configured with DocumentRoot and available to CGIs via the DOCUMENT_ROOT environment variable). In a normal configuration, this is used by the core module when mapping URIs to filenames, but when the server is configured to do dynamic virtual hosting, that job must be taken over by another module (either mod_vhost_alias or mod_rewrite), which has a different way of doing the mapping. Neither of these modules is responsible for setting the DOCUMENT_ROOT environment variable so if any CGIs or SSI documents make use of it, they will get a misleading value.

Simple Dynamic Virtual Hosts

This extract from httpd.conf implements the virtual host arrangement outlined in the Motivation section above, but in a generic fashion using mod_vhost_alias.

This configuration can be changed into an IP-based virtual hosting solution by just turning UseCanonicalName Off into UseCanonicalName DNS. The server name that is inserted into the filename is then derived from the IP address of the virtual host.

A Virtually Hosted Homepages System

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, for example, the documents for www.user.isp.com are found in /home/user/. It uses a single cgi-bin directory instead of one per virtual host.

There are examples of more complicated VirtualDocumentRoot settings in the mod_vhost_alias documentation.

Using Multiple Virtual Hosting Systems on the Same Server

With more complicated setups, you can use Apache's normal <VirtualHost> directives to control the scope of the various virtual hosting configurations. For example, you could have one IP address for general customers' homepages, and another for commercial customers, with the following setup. This can, of course, be combined with conventional <VirtualHost> configuration sections.

More Efficient IP-Based Virtual Hosting

The configuration changes suggested to turn the first example into an IP-based virtual hosting setup result in a rather inefficient setup. A new DNS lookup is required for every request. To avoid this overhead, the filesystem can be arranged to correspond to the IP addresses, instead of to the host names, thereby negating the need for a DNS lookup. Logging will also have to be adjusted to fit this system.

Simple Dynamic Virtual Hosts Using mod_rewrite

This extract from httpd.conf does the same thing as the first example. The first half is very similar to the corresponding part above, except for some changes, required for backward compatibility and to make the mod_rewrite part work properly; the second half configures mod_rewrite to do the actual work.

There are a couple of especially tricky bits: by default, mod_rewrite runs before other URI translation modules (mod_alias etc.) - so if you wish to use these modules, mod_rewrite must be configured to accommodate them. Also, some magic is required to do a per-dynamic-virtual-host equivalent of ScriptAlias.

A Homepages System Using mod_rewrite

This does the same thing as the second example.

Using a Separate Virtual Host Configuration File

This arrangement uses more advanced mod_rewrite features to work out the translation from virtual host to document root, from a separate configuration file. This provides more flexibility, but requires more complicated configuration.

The vhost.map file should look something like this:

The httpd.conf should contain the following: