Mailing-List: contact users-help@httpd.apache.org; run by ezmlm
Precedence: bulk
Reply-To: users@httpd.apache.org
Received-SPF: pass (nike.apache.org: domain of tevans.uk@googlemail.com
 designates 209.85.212.45 as permitted sender)
MIME-Version: 1.0
In-Reply-To: 
 <0ACABF6B29D1F84F99B2BB7866923B0F51CA6189@0303mail03.cecred.coop.br>
References: 
 <0ACABF6B29D1F84F99B2BB7866923B0F51CA6189@0303mail03.cecred.coop.br>
Date: Thu, 15 Dec 2011 12:41:44 +0000
Message-ID: 
 <CAFHbX1LEhoWmrAKyQONGvSK-c3cXtDvB0cei0WVUGm3F0TdiNg@mail.gmail.com>
From: Tom Evans <tevans.uk@googlemail.com>
To: users@httpd.apache.org
Content-Type: text/plain; charset=UTF-8
Subject: Re: [users@httpd] high load average

On Thu, Dec 15, 2011 at 12:18 PM, Diego Maciel Gomes
<diego.gomes@cecred.coop.br> wrote:
> hey Guys...
>
> I need help :)
>
> I have an intranet running with LAMP. I have a server with 8gb mem and 4
> procs for this. My apache is 2.2.3
>
> I have about 400 users connected (ps -ef |grep httpd |wc -l)
>
> At this moment, Im using 4.3gb of mem and sometimes I have a high use of
> procs..
>
> My load average at this moment is 4.62, 4.77 and 4.32. more than three times
> a day, I have a high use of average, like 34.15, 23.71 and 16.89.
>
> Im thinking to do an apache cluster.. but, I dont know if it is the correct
> way...
>
> I dont have any change in my intranet more than 6 months... I feel that this
> high load average started when I did an apache update...
>
> Anyone have any idea to help me?
>
> Thanks anyway,
>
> Diego
>

You updated to Apache 2.2.3? That is an extremely old version - we're
now on 2.2.21 - what version did you start from :o

It almost certainly is not Apache that is the issue. Apache is
probably the most simple and consistent part of a LAMP stack, and
fetching and serving content is not a computationally expensive task.

A high load average means you have lots of processes on the run
queue/waiting for IO, on a LAMP stack this probably means you have
many requests waiting for their turn at the database, which results in
requests taking longer, and that means you will have more concurrent
requests, which means more Apache processes/threads/workers/memory.

More concurrent requests will usually also mean more concurrent
queries to the database, meaning things only get worse until the
number of users accessing the site dies down.

LAMP is not one thing, it is a stack. Therefore you should look at all
aspects of the stack:

PHP memory usage can be an issue. You should check how much memory and
CPU PHP is using by serving it in a more efficient manner, eg using
mod_fcgid, so that it is isolated from Apache.
MySQL gets slow when overloaded. Are queries taking too long in peak
times? Do you need to scale out the DB? Are writes destroying read
speed?

Cheers

Tom

---------------------------------------------------------------------
The official User-To-User support forum of the Apache HTTP Server Project.
See <URL:http://httpd.apache.org/userslist.html> for more info.
To unsubscribe, e-mail: users-unsubscribe@httpd.apache.org
   "   from the digest: users-digest-unsubscribe@httpd.apache.org
For additional commands, e-mail: users-help@httpd.apache.org