httpd-users mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Joshua Slive" <>
Subject Re: [users@httpd] performance considerations (looong)
Date Tue, 18 Jul 2006 21:36:18 GMT
On 7/18/06, Pavel Stratil <> wrote:
> Hi all!
> I have an older machine with no real funds to upgrade hw and growing
> traffic. Apache eats most of the resources, before some basic config
> tweaking enough to take the machine down (now average cpu load 35% (peak
> 70), avg mem usage 85% (peak 100)). The biggest issue is memory and
> processor load and also traffic (about 1 TB/month). To be able to
> further optimize the performance, i'd rely on your experiences. I ask a
> lot at once so please dont be afraid of partial answers:-)

As a general comment, this question is far too broad to get a decent
response on a forum like this.  This is more like something you would
drop in the hands of a consultant.

If you want help from the list, I'd break your issue into small,
specific questions.  Then there is a better hope of getting
constructive responses.

To hit a few of your specific questions:

- DYNAMIC_MODULE_LIMIT isn't going to help you in any substantial way.
 I'd just ignore it.

- Caching is documented here:
You should really use 2.2, and you should start with mod_disk_cache as
the most generally useful solution.  I believe that, in the case of
mod_deflate-processed content, the cache will wind up storing both a
deflated and a non-deflated variant since both will be requested by

Caching could help you alot, but your dynamic content needs to be
carefully structured to allow for caching.  If high CPU usage is the
problem, you should spend most of your time optimizing your php
scripts.  Any change you make in apache will just be noise compared to
that.  For memory, your idea of splitting into dynamic and static
servers is good.  Both servers should be stripped to the bare minimum


The official User-To-User support forum of the Apache HTTP Server Project.
See <URL:> for more info.
To unsubscribe, e-mail:
   "   from the digest:
For additional commands, e-mail:

View raw message