Mailing-List: contact users-help@httpd.apache.org; run by ezmlm
Precedence: bulk
Reply-To: users@httpd.apache.org
Message-ID: <40089A5E.9020709@mindspring.com>
Date: Fri, 16 Jan 2004 21:13:50 -0500
From: Aaron W Morris <aaronmorris@mindspring.com>
User-Agent: Mozilla/5.0 (Windows; U; Windows NT 5.1; en-US;
 rv:1.5) Gecko/20031007
MIME-Version: 1.0
To: users@httpd.apache.org
References: <40079471.8050705@damianmeyer.com>
In-Reply-To: <40079471.8050705@damianmeyer.com>
Content-Type: text/plain; charset=us-ascii; format=flowed
Content-Transfer-Encoding: 7bit
Subject: Re: [users@httpd] Timeouts / responsiveness issues with Apache

Damian Meyer wrote:
> Hi,
> 
> We have recently migrated from several individual Apache 1.3.x servers 
> running around 1000 virtual sites each to a server farm with multiple 
> servers (Apache/1.3.29) all capable of running all of the sites (about 
> 4000 in total).  The server farm is behind a load balancer running 
> ldirectord from the LVS project.
> 
> The problem we are having is that every so often one of the web servers 
> seems to go offline and either become very sluggish with responses or 
> even times-out.  If it failed totally, ldirectord would notice and map 
> it out, but it doesn't (usually).  After a few minutes of bad 
> performance, it usually comes good, but it happens frequently enough to 
> one or more of the web servers that it's causes noticeable glitches to 
> our clients.
> 
> Also, and it may or may not be related, occasionally Apache dies totally 
> (ie no live processes) and requires starting again.
> 
> Each server has 2GB of RAM.  Websites are a mix of flat HTML through to 
> complicated PHP sites using MySQL connections (we've checked the 
> database, it doesn't look like this is the problem.)
> 
> Does anyone have any idea what could be at issue here?  Does anyone have 
> any experience with this sort of large-ish server farm to give us any 
> feedback on configuration settings or other tuning parameters?  Any 
> decent tuning documents on the web that take into account this type of 
> scenario?
> 
> Kind regards
> Damian Meyer

Since you are using PHP [with many sites], I would consider lowering 
"MaxRequestsPerChild" so that child processes die after a reasonable 
amount of requests.

or

I would consider trying regular restarts of apache via cron.  Since you 
have load balanced servers, you should not be too concerned with it 
failing to restart on one server.  I like Apache and all, but I have had 
issues with process "build-up" on heavily used instances.


-- 
Aaron W Morris <aaronmorris@mindspring.com> (decep)


---------------------------------------------------------------------
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