Advanced Techniques with mod_rewrite
This document supplements the mod_rewrite
reference documentation. It provides a few advanced techniques using mod_rewrite.
URL-based sharding across multiple backends
- Description:
-
A common technique for distributing the burden of server load or storage space is called "sharding". When using this method, a front-end server will use the url to consistently "shard" users or objects to separate backend servers.
- Solution:
-
A mapping is maintained, from users to target servers, in external map files. They look like:
user1 physical_host_of_user1 user2 physical_host_of_user2 : :
We put this into a
map.users-to-hosts
file. The aim is to map;/u/user1/anypath
to
http://physical_host_of_user1/u/user/anypath
thus every URL path need not be valid on every backend physical host. The following ruleset does this for us with the help of the map files assuming that server0 is a default server which will be used if a user has no entry in the map:
RewriteEngine on RewriteMap users-to-hosts "txt:/path/to/map.users-to-hosts" RewriteRule "^/u/([^/]+)/?(.*)" "http://${users-to-hosts:$1|server0}/u/$1/$2"
See the RewriteMap
documentation for more discussion of the syntax of this directive.
On-the-fly Content-Regeneration
- Description:
-
We wish to dynamically generate content, but store it statically once it is generated. This rule will check for the existence of the static file, and if it's not there, generate it. The static files can be removed periodically, if desired (say, via cron) and will be regenerated on demand.
- Solution:
- This is done via the following ruleset:
# This example is valid in per-directory context only RewriteCond "%{REQUEST_URI}" "!-U" RewriteRule "^(.+)\.html$" "/regenerate_page.cgi" [PT,L]
The
-U
operator determines whether the test string (in this case,REQUEST_URI
) is a valid URL. It does this via a subrequest. In the event that this subrequest fails - that is, the requested resource doesn't exist - this rule invokes the CGI program/regenerate_page.cgi
, which generates the requested resource and saves it into the document directory, so that the next time it is requested, a static copy can be served.In this way, documents that are infrequently updated can be served in static form. if documents need to be refreshed, they can be deleted from the document directory, and they will then be regenerated the next time they are requested.
Load Balancing
- Description:
-
We wish to randomly distribute load across several servers using mod_rewrite.
- Solution:
-
We'll use
RewriteMap
and a list of servers to accomplish this.RewriteEngine on RewriteMap lb "rnd:/path/to/serverlist.txt" RewriteRule "^/(.*)" "http://${lb:servers}/$1" [P,L]
serverlist.txt
will contain a list of the servers:## serverlist.txt servers one.example.com|two.example.com|three.example.com
If you want one particular server to get more of the load than the others, add it more times to the list.
- Discussion
-
Apache comes with a load-balancing module -
mod_proxy_balancer
- which is far more flexible and featureful than anything you can cobble together using mod_rewrite.
Structured Userdirs
- Description:
-
Some sites with thousands of users use a structured homedir layout, i.e. each homedir is in a subdirectory which begins (for instance) with the first character of the username. So,
/~larry/anypath
is/home/l/larry/public_html/anypath
while/~waldo/anypath
is/home/w/waldo/public_html/anypath
. - Solution:
-
We use the following ruleset to expand the tilde URLs into the above layout.
RewriteEngine on RewriteRule "^/~(([a-z])[a-z0-9]+)(.*)" "/home/$2/$1/public_html$3"
Redirecting Anchors
- Description:
-
By default, redirecting to an HTML anchor doesn't work, because mod_rewrite escapes the
#
character, turning it into%23
. This, in turn, breaks the redirection. - Solution:
-
Use the
[NE]
flag on theRewriteRule
. NE stands for No Escape. - Discussion:
- This technique will of course also work with other special characters that mod_rewrite, by default, URL-encodes.
Time-Dependent Rewriting
- Description:
-
We wish to use mod_rewrite to serve different content based on the time of day.
- Solution:
-
There are a lot of variables named
TIME_xxx
for rewrite conditions. In conjunction with the special lexicographic comparison patterns<STRING
,>STRING
and=STRING
we can do time-dependent redirects:RewriteEngine on RewriteCond "%{TIME_HOUR}%{TIME_MIN}" ">0700" RewriteCond "%{TIME_HOUR}%{TIME_MIN}" "<1900" RewriteRule "^foo\.html$" "foo.day.html" [L] RewriteRule "^foo\.html$" "foo.night.html"
This provides the content of
foo.day.html
under the URLfoo.html
from07:01-18:59
and at the remaining time the contents offoo.night.html
.mod_cache
, intermediate proxies and browsers may each cache responses and cause the either page to be shown outside of the time-window configured.mod_expires
may be used to control this effect. You are, of course, much better off simply serving the content dynamically, and customizing it based on the time of day.
Set Environment Variables Based On URL Parts
- Description:
-
At time, we want to maintain some kind of status when we perform a rewrite. For example, you want to make a note that you've done that rewrite, so that you can check later to see if a request can via that rewrite. One way to do this is by setting an environment variable.
- Solution:
-
Use the [E] flag to set an environment variable.
RewriteEngine on RewriteRule "^/horse/(.*)" "/pony/$1" [E=rewritten:1]
Later in your ruleset you might check for this environment variable using a RewriteCond:
RewriteCond "%{ENV:rewritten}" "=1"
Note that environment variables do not survive an external redirect. You might consider using the [CO] flag to set a cookie.
Please login to continue.