Gentoo Forums
Gentoo Forums
Gentoo Forums
Quick Search: in
Apache procs creating abnormal high load
View unanswered posts
View posts from last 24 hours

 
Reply to topic    Gentoo Forums Forum Index Networking & Security
View previous topic :: View next topic  
Author Message
linkmark
n00b
n00b


Joined: 15 Nov 2004
Posts: 36
Location: belgium

PostPosted: Tue Jun 27, 2006 12:46 pm    Post subject: Apache procs creating abnormal high load Reply with quote

Hi all.

I'm running an webserver for a small organization who rents webspace to people we know.
But the last few day's apache is acting very strange, there are some procs who are running each on the same cpu % load whyle the server should be almost idle. Resulting in constantly 100% load. I can't find the reason for this, searched google, this forum, ..

Here is what i get with top:
Code:
Tasks: 139 total,   7 running, 132 sleeping,   0 stopped,   0 zombie
Cpu(s): 100.0% us,  0.0% sy,  0.0% ni,  0.0% id,  0.0% wa,  0.0% hi,  0.0% si
Mem:    905736k total,   895328k used,    10408k free,   127776k buffers
Swap:   787176k total,    41076k used,   746100k free,   557824k cached

  PID USER      PR  NI  VIRT  RES  SHR S %CPU %MEM    TIME+  COMMAND                                   
29415 apache    25   0  101m  39m 3424 R 19.9  4.4  10:04.15 apache2                                     
29424 apache    25   0  101m  39m 3424 R 19.9  4.4  13:38.45 apache2                                     
29432 apache    25   0  125m  62m 3340 R 19.9  7.0  22:46.14 apache2                                     
29454 apache    25   0  125m  62m 3340 R 19.9  7.0  30:28.91 apache2                                     
29413 apache    25   0  125m  62m 3340 R 19.6  7.0  17:30.76 apache2                                     
 2228 root      16   0  1976 1108  828 R  0.3  0.1   0:00.02 top                                         
    1 root      16   0  1428  280  248 S  0.0  0.0   0:31.62 init
..


In the apache error log i find things like this:
Code:
[Thu Jun 15 08:24:38 2006] [error] [client 216.39.33.185] client sent HTTP/1.1 request without hostname (see RFC2616 section 14.23): /w00tw00t.at.ISC.SANS.DFind:)
[Thu Jun 15 11:51:00 2006] [notice] SIGUSR1 received.  Doing graceful restart
[Thu Jun 15 11:51:00 2006] [notice] Digest: generating secret for digest authentication ...
[Thu Jun 15 11:51:00 2006] [notice] Digest: done
[Thu Jun 15 11:51:01 2006] [notice] Apache configured -- resuming normal operations
[Thu Jun 15 11:51:01 2006] [notice] SIGUSR1 received.  Doing graceful restart
[Thu Jun 15 11:51:01 2006] [notice] Digest: generating secret for digest authentication ...
[Thu Jun 15 11:51:01 2006] [notice] Digest: done
[Thu Jun 15 11:51:01 2006] [notice] Apache configured -- resuming normal operations
[Thu Jun 15 11:51:02 2006] [notice] SIGUSR1 received.  Doing graceful restart
[Thu Jun 15 11:51:02 2006] [notice] Digest: generating secret for digest authentication ...
[Thu Jun 15 11:51:02 2006] [notice] Digest: done
[Thu Jun 15 11:51:03 2006] [notice] Apache configured -- resuming normal operations
[Thu Jun 15 11:51:03 2006] [notice] SIGUSR1 received.  Doing graceful restart
[Thu Jun 15 11:51:03 2006] [notice] Digest: generating secret for digest authentication ...
[Thu Jun 15 11:51:03 2006] [notice] Digest: done
..


and (A LOT OF THESE, almost the whole log)

Code:
sh: /which: No such file or directory
sh: /which: No such file or directory
sh: /which: No such file or directory
sh: /which: No such file or directory
sh: /which: No such file or directory
sh: /which: No such file or directory
sh: /which: No such file or directory
sh: /which: No such file or directory
sh: /which: No such file or directory
sh: /which: No such file or directory
..


and

Code:
[Tue Jun 27 01:15:48 2006] [warn] child process 24124 still did not exit, sending a SIGTERM
[Tue Jun 27 01:15:48 2006] [warn] child process 24126 still did not exit, sending a SIGTERM
[Tue Jun 27 01:15:50 2006] [warn] child process 24124 still did not exit, sending a SIGTERM
[Tue Jun 27 01:15:50 2006] [warn] child process 24126 still did not exit, sending a SIGTERM
[Tue Jun 27 01:15:52 2006] [warn] child process 24124 still did not exit, sending a SIGTERM
[Tue Jun 27 01:15:52 2006] [warn] child process 24126 still did not exit, sending a SIGTERM
[Tue Jun 27 01:15:54 2006] [error] child process 24124 still did not exit, sending a SIGKILL
[Tue Jun 27 01:15:54 2006] [error] child process 24126 still did not exit, sending a SIGKILL
[Tue Jun 27 01:15:55 2006] [notice] caught SIGTERM, shutting down
[Tue Jun 27 01:16:12 2006] [notice] suEXEC mechanism enabled (wrapper: /usr/sbin/suexec2)
[Tue Jun 27 01:16:13 2006] [notice] Digest: generating secret for digest authentication ...
[Tue Jun 27 01:16:13 2006] [notice] Digest: done
[Tue Jun 27 01:16:13 2006] [notice] Apache configured -- resuming normal operations
..


This last error messages ocure when i restart apache (what gives me no problem, the restarting). Afther the restart of apache, he run's smooth again for a whyle, until, for i don't know what reason yet, those loaded proc's are back.

Anyone has any idea? Would help me a lot.
Thanks in advanced!

linkmark
Back to top
View user's profile Send private message
celestialwizard
Tux's lil' helper
Tux's lil' helper


Joined: 15 Jun 2006
Posts: 81
Location: Brisbane, Australia

PostPosted: Wed Jun 28, 2006 4:41 am    Post subject: Reply with quote

what version of apache are you running?
what are your virtual servers actually doing? (php? (4? 5?), phython?, perl?, java?, etc...)

why do you have suExec enabled? - disable it if possible, and if not possible, make it so.
Back to top
View user's profile Send private message
selig
Guru
Guru


Joined: 31 Jul 2005
Posts: 425
Location: Prague, Czech Republic

PostPosted: Wed Jun 28, 2006 9:52 am    Post subject: Reply with quote

I guess I am having a similar problem. Apache runs fine, but every once in a few days, tons of apache processes become active and eat up the CPU. As a result of this, the load goes up to 60 or more and apache slows down to unusable levels, I have to stop it and start it again afterwards. I am using version 2.0.55-r1:

Code:

 net-www/apache-2.0.55-r1  USE="apache2 mpm-prefork ssl threads -debug -doc -ldap -mpm-leader -mpm-peruser -mpm-threadpool -mpm-worker -no-suexec -static-modules"



This is my setting from httpd.conf - maybe I should limit the MaxClients directive?:

Code:

<IfModule prefork.c>
StartServers         5
MinSpareServers      5
MaxSpareServers     10
MaxClients         150
MaxRequestsPerChild 1200
</IfModule>


I run nothing special on it, just PHP with mod_php, but even serving static pages gets slow after that happens. Oh yes, the server gets 10 requests per second on average, so I am afraid lowering hte MaxClients directive significantly will cause decreased performance.


Last edited by selig on Wed Jun 28, 2006 2:55 pm; edited 1 time in total
Back to top
View user's profile Send private message
linkmark
n00b
n00b


Joined: 15 Nov 2004
Posts: 36
Location: belgium

PostPosted: Wed Jun 28, 2006 2:16 pm    Post subject: Reply with quote

celestialwizard wrote:
what version of apache are you running?
what are your virtual servers actually doing? (php? (4? 5?), phython?, perl?, java?, etc...)

why do you have suExec enabled? - disable it if possible, and if not possible, make it so.

My version: apache-2.0.55
The virtual hosts are created by Syscp ( http://www.syscp.de/ ) and each virtual hosts is an user account. The only thing that is enabled for these virtual hosts is PHP5 ( Version 5.1.1-gentoo )

Example
Code:
# Domain ID: 100 - CustomerID: 45 - CustomerLogin: dries
<VirtualHost 86.39.128.150>
  ServerName chiromais.be
  ServerAlias www.chiromais.be
  ServerAdmin driesweckx@gmail.com
  DocumentRoot /srv/www/users/dries/
  php_admin_flag safe_mode On
  ErrorLog /srv/www/logs/dries-error.log
  CustomLog /srv/www/logs/dries-access.log combined


suExec is disabled now, but this isn't the reason for the load right?
Back to top
View user's profile Send private message
celestialwizard
Tux's lil' helper
Tux's lil' helper


Joined: 15 Jun 2006
Posts: 81
Location: Brisbane, Australia

PostPosted: Wed Jun 28, 2006 11:07 pm    Post subject: Reply with quote

To provide a quick look into what Apache is doing, enable service-status for a trusted requester and check it under normal load and under high load.

that should give you more of an idea of what process is running hot, who from, etc...

much easier than correlating logs and process output
Back to top
View user's profile Send private message
linkmark
n00b
n00b


Joined: 15 Nov 2004
Posts: 36
Location: belgium

PostPosted: Sun Feb 11, 2007 3:42 pm    Post subject: Reply with quote

quite strange, the high load stopped itself .. without me doing something ..
Back to top
View user's profile Send private message
Display posts from previous:   
Reply to topic    Gentoo Forums Forum Index Networking & Security All times are GMT
Page 1 of 1

 
Jump to:  
You cannot post new topics in this forum
You cannot reply to topics in this forum
You cannot edit your posts in this forum
You cannot delete your posts in this forum
You cannot vote in polls in this forum