httpd-users mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Tom Evans <>
Subject Re: [users@httpd] Apache 2.2 optimization -> Three tips
Date Mon, 22 Mar 2010 15:55:31 GMT
On Mon, Mar 22, 2010 at 3:45 PM, Geoff Millikan <> wrote:
>> Fair enough, but if your testing was of ~30 requests, and we
>> are believing that
>> the typical browser is making 6 simultaneous connections,
>> then it sounds like
>> the real magic was 6 * fudge factor of 5 ;-P
> LOL, yes, I totally agree and I kept thinking that no optimization guide has
> MinSpareServers and StartServers in the 20ies, 30ies or 40ies so I kept
> telling myself that there's no way that the "waiting" issues we saw could be
> solved by increasing the Servers.  However, we eliminated everything else:
> we're on 100Mbit Internet connection in 1 Wilshire, Los Angeles which is one
> of the more interconnected buildings in the USA, network optimized (all
> hardware Cisco GigE rated), server NIC cards all GigE rated and tuned, file
> system optimized (see noatime) and disk buffers optimized on RAID 10, extra
> RAM to spare, DNS resolution times tested, etc, etc, etc.  We were going
> crazy. We started looking into serving content with Lighttpd or Nginx to try
> and get around the issue but didn't want to configure/maintain more web
> servers so went back to the drawing board with Apache.
> Increasing MinSpareServers and StartServers beyond "normal" is what solved
> the issue for us.
> Although it doesn't make sense it seems 5 or 10 or more servers cannot
> *simultaneously* both read 35 files off the server and serve them.  Now if
> the Client is on dial-up, or mediocre DSL then the extra 500ms to 800ms of
> wait time probably wouldn't even be noticed as the last few images streaming
> in and the page finishes rendering.  But our office is on a 20 Mbps Verizon
> fiber link, and many of our customers are on fast pipes.  So we really do
> notice the difference between a web page that "snaps up to attention" and
> one that renders 90% with the last few pieces shuffling around.
> Best,
> -+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
> "The nine most terrifying words in the English language are: I'm from the
> government and I'm here to help." - Ronald Reagan, August 21, 1986
> -+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+

Have you considered using a more resource effective MPM, like worker
or event? (disclaimer: I love the event MPM!)

If it is pure static hosting, then worker/event is a no-brainer. For
PHP/Python/etc you simply run the web app out of process and
communicate via WSGI or FastCGI (WSGI if you just do Python, FastCGI
if you prefer to be language agnostic).



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