After updating from Lenny to Squeeze, my Apache2 server has been acting
really weird. Hundreds of pids will have been spawned, filling up RAM
and filling up swap, causing the kernel OOM killer to start. Nothing in
/var/log/messages or dmesg gives any indication of what is happening.
Nothing in /var/log/apache2/access.log or error.log either that could be
of help.
This isn't a robust server, so when it happens, it brings the server to
its knees, and it's down for hours before the kernel manages to kill the
pids and restore sanity. Further, there doesn't seem to be any standard
time interval when the pids are created. I've seen the server go 6 hours
after boot before the problems start, and I've seen it go as long as 3
days. It seems entirely arbitrary on when it's going to happen.
I've got kernels 2.6.32-5-686, 2.6.26-2-686 and 2.6.26-1-686 installed.
The problem exists with all three booted kernels. Apache2 version is
2.2.16-6.
<IfModule mpm_prefork_module>
StartServers 5
MinSpareServers 5
MaxSpareServers 10
MaxClients 150
MaxRequestsPerChild 0
</IfModule>
I am not aware of any changes that I have made to the apache2.conf since
the upgrade, nor am I aware of any changes that dpkg made during the
upgrade.
I don't want to reinstall the server unless I absolutely have to. I do
have all my data backed up, so it's not a concern of doing so, it's just
that I don't have the time to dedicate to reinstalling and restoring
data.
What can I do to troubleshoot this problem, and restore sanity to my
Apache installation? Any help will be greatly appreciated.
--
. o . o . o . . o o . . . o .
. . o . o o o . o . o o . . o
o o o . o . . o o o o . o o o
Attachment:
signature.asc
Description: Digital signature