Hi Mike,
Well, the above LAMP configuration is way under this, built from
dust-covered parts picked out from the trash, and it's hard to find 2 gig
memory in the trash... :)
About the main question, having 250 different logfiles is not a number
that a usual webserver would cry about, even if it has to write 4-5
entries per sec into each of them. However, if you have all these entries
in one file, it would take much extra time to split them, analyze them,
update the scripts that does the job, etc... I'm too comfortable to do
this, this would not be an option, even if I had 2000 vhosts.
Apart from certain usage - like MP3 or video streaming on a 100 or
1000Mbps line -, the bottleneck is not the disk subsystem, but the CPU is,
the average load of 2 also shows this. Using a RAID1 array also decreases
stress on the disks, the 4.5% iowait avg is not an issue - and logging
into different logfiles does not increase this at all. You should rather
look for a faster CPU, because every other part of your box seems to be
fine - at least from my point of view.
Regards,
Zoltan HERPAI
On Wed, 31 May 2006, Mike Jackson wrote:
I'm running a couple Redhat servers with Apache 1.3.36 (which my company is
sticking with for legacy reasons; I'm sure we'll move to 2.0 or 2.2
eventually, but everyone here other than me has only used 1.3). It's a
typical LAMP setup. There's about 250 virtual hosts on a handful of IPs on
the server I'm particularly concerned with. It averages about 13 busy Apache
children, bursts up to 40 on a typical day, and MaxChildren is set to 60. The
system load is typically over 2, and the bottleneck seems to be the disk
(which is a two-disk SCSI RAID 1 array). iowait is around 4.5%. There's 2GB
RAM, which is pretty heavily used (it averaged 75MB free over the last 24
hours). The swap partition is on a secondary IDE drive, and receives minimal
use.
We currently log all the virtual host activity to individual log files. Would
it be more efficient to log to single logfile that's later split for each
vhost? Or would that make the disk utilization worse? Or should I be looking
to improve efficiency somewhere else? I already know the site PHP code is a
bit inefficient in how it uses the disks, and I've talked to the programmer
about cleaning that up a bit.
---------------------------------------------------------------------
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: [EMAIL PROTECTED]
" from the digest: [EMAIL PROTECTED]
For additional commands, e-mail: [EMAIL PROTECTED]
---------------------------------------------------------------------
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: [EMAIL PROTECTED]
" from the digest: [EMAIL PROTECTED]
For additional commands, e-mail: [EMAIL PROTECTED]
