On Mon, 2 Oct 2006, Devraj Mukherjee wrote:
On 10/2/06, Devraj Mukherjee devraj@gmail.com wrote:
On 10/2/06, Dag Wieers dag@wieers.com wrote:
On Mon, 2 Oct 2006, Devraj Mukherjee wrote:
I suspect that Apache 2.0.52 as distributed with CentOS has a memory leak issue. Is anyone aware of this?
Being more specific might make your mail more valuable. A memory leak in what module ? How to reproduce it ? Is it a known Apache issue or something specific to how the CentOS Apache has been compiled ? Why do you suspect there is a memory leak ?
Any reference would make your post useful, now it's not.
BTW Did you know the kernel you are running has a bug ?
you mean 2.6.9-42.0.2.plus.c4smp
It doesn't matter what you actually are running. I was just making the point that without specifics the statement is fairly useless to most.
However the information below is just a start.
Also notice the following messages
Oct 1 07:49:47 charlie kernel: Out of Memory: Killed process 1996 (httpd). Oct 1 07:49:47 charlie kernel: oom-killer: gfp_mask=0x1d2
[snip]
Oct 1 07:49:48 charlie kernel: Out of Memory: Killed process 31946 (httpd).
What we now need is the context.
Was it really httpd that was causing it ? What were you doing when it happened ? What technologies are you using that may cause this ? Did you try another Apache (a more recent build) ? Did you try to unload unnecessary modules to find the culprit ?
Everything you think is useful to color in the picture and to give a fairly good idea what you have set up and what you think might have caused it.
There are many reasons why Apache could use up memory and it's even, possible that a configuration error could lead up to this.
To find the cause, much more information is required.
Then again, I don't think the CentOS mailinglist is the best location to find the cause. The first thing you probably need to know is: is this a known issue with Apache 2.0 or did you find a new problem (possible related to some other software). For that you need to look for similar incidents on the Apache forums/bugzilla.
If you require support, you may want to buy a Red Hat Enterprise version, test it on that and in case you still have the problem, open a ticket with Red Hat.
Whatever the cause, an action has to come from you depending how you want it solved. Moving to Apache 2.2 may solve the problem at hand but may not be the best problem overall.
Kind regards, -- dag wieers, dag@wieers.com, http://dag.wieers.com/ -- [all I want is a warm bed and a kind word and unlimited power]