I have also noticed kswapd daemon causing the load on server. What i believe is that there is some website whose pages uses lot of memory. Linux uses kswapd for virtual memory management such that pages that have been recently accessed are kept in memory and less active pages are paged out to disk. So,the kswapd process regularly decreases the ages of unreferenced pages...and at the end they are paged out(moved out) to disk.
I am using Apache 2.6. My question is that is it possible to find out which website or script causing this? I used apachetop and find a website having lot of hits & consuming lot of bandwidth. After suspending it, still the issue is there. Top command is showing only httpd processes.
I have tried everything mentioned (mod_evasive,mod_cband,etc) but still clueless. Is it possible in apache2 to run each website under different userid? If yes then how it can be achieved? This will probably ease out to find the culprit.
Thanks,
Gaurav Pruthi--On Fri, Aug 29, 2008 at 2:53 PM, Alexandru David Constantinescu <aldavx@xxxxxxxxx> wrote:
Gaurav Pruthi wrote:
Hi Agnello,
I am using prefork MPM
<IfModule mpm_prefork_module>
StartServers 20
MinSpareServers 25
MaxSpareServers 50
MaxClients 500
MaxRequestsPerChild 0
</IfModule>
Thanks,
Gaurav
On Fri, Aug 29, 2008 at 1:30 PM, Agnello George <agnello.dsouza@xxxxxxxxx <mailto:agnello.dsouza@xxxxxxxxx>> wrote:
On 8/29/08, *Gaurav Pruthi* <gkpruthi@xxxxxxxxx<mailto:gkpruthi@xxxxxxxxx>> wrote:www.linux-vashi.blogspot.com <http://www.linux-vashi.blogspot.com> <http://www.bible-study-india.blogspot.com>
Dear Friends,
I am facing very strange problem. I have a LAMP server. It has
around 400 of websites hosted on it. Since last one month the
server load increases suddenly. When checked using top command,
i saw apache process consuming 99% of CPU. After i kill httpd
and starts it again (or using HUP), load comes down to the
normal. It happens 8-10 times each day. I am unable to find what
could cause the apache to consume all the resources of server
and causing server to hang.
Kindly help me to resolve this issue.
Could you tell us the Server-Pool Size Regulation mentioned in the
httpd.conf -- Regards
Agnello D'souza
--
Gaurav Pruthi
MaxClients 500 could be useless (I may be wrong but from what I know is a hard coded limit in apache source code HARD_SERVER_LIMIT who is set to 250 so if you want to raise that limit you must recompile your apache)
MaxRequestsPerChild 0 it say that once a child is created he can stay up and running for infinite time. It could be better to limit this , let say to 10000 in order to force apache to kill a stressed child and start a fresh one.
MinSpareServers 25 in my opinion is a little bit high. Practically you force apache to keep free 25 servers any time. Try to set this parameter to something like 10 or 15
Ther is a lot of literature about those parameters and how is the best way to set them bun ther is not a clear solutions so the best practice is to start playing with them and look for what happen.
---------------------------------------------------------------------
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@xxxxxxxxxxxxxxxx
" from the digest: users-digest-unsubscribe@xxxxxxxxxxxxxxxx
For additional commands, e-mail: users-help@xxxxxxxxxxxxxxxx
Gaurav Pruthi
|