Apache segmentation fault

faris · Unread post by **faris** » Sun Oct 19, 2008 3:09 pm

Hmm... but thinking about it, Scott has httpd-debuginfo installed and he was still getting segfaults.

However, I think it may be related to moving bits of memory around, and the amount of memory installed and stuff like that.

Faris.

hostingguy · Unread post by **hostingguy** » Mon Nov 10, 2008 5:30 pm

aus-city wrote:Its STILL behaving

This is the only time I can remember where some one reported that something was working with a frown :p

Unread post by **scott** » Mon Nov 10, 2008 7:57 pm

Haha

we were trying to get a repeatable environment of it crashing so we could test fixes.

faris · Unread post by **faris** » Tue Nov 11, 2008 8:39 am

I think I can add a little more info now, though in a relatively non-technical and somewhat "guessful" way.

From my monitoring of memory usage and segfaults, it would appear that one possible cause of the problem is that apache is failing to release memory it no longer needs, or at any rate is getting confused when it comes to memory.

Where this comes into the equation, and why adding the debug stuff solves the problem for some people is beyond my knowledge.

Faris.

aus-city · Unread post by **aus-city** » Mon Nov 17, 2008 2:06 pm

faris,

Its still behaving with the debug running - go figure it must be changing the memory enough its not happening anymore.

hostingguy · Unread post by **hostingguy** » Mon Nov 17, 2008 3:00 pm

Are you guys using the httpd-2.2.3-11.el5_2.centos.4 build ?

faris · Unread post by **faris** » Mon Nov 17, 2008 3:06 pm

No.

I'm on httpd-2.0.63-2.el4s1.centos.2 for these particular ones.

I think this is from the centosplus repo (for centos 4)

Faris.

hostingguy · Unread post by **hostingguy** » Mon Nov 17, 2008 4:21 pm

are you able to use the 2.2.x branch on centos/rhel 4 or are you limited to 2.0 ?

is the problem only occuring on 2.0 ?

Unread post by **scott** » Mon Nov 17, 2008 8:55 pm

neat, I hadn't been tracking the 2.2.x packages for C4. Does that work with Plesk?

faris · Unread post by **faris** » Wed Nov 19, 2008 10:06 am

We are starting to see segfaults again, but only occasionally. I had put a freeze on updating the rules since this started, but recently did some updates.

Although I'm still disabling the domain-blacklists I guess the other rules files have increased in size or complexity since I last updated.

Over the weekend I'm going to try adding the debug thing to see if that helps.

There also seems to be a new version of httpd for centos. Who knows, that may also help.

Faris.

hostingguy · Unread post by **hostingguy** » Thu Nov 20, 2008 2:56 pm

scott wrote:neat, I hadn't been tracking the 2.2.x packages for C4. Does that work with Plesk?

Seems to work ok. I get only a couple of seg faults across a lot of servers and am not sure it even applies to this situation.

Unread post by **scott** » Fri Nov 21, 2008 11:07 am

Thats awesome, where did that come from, centosplus?

Unread post by **mikeshinn** » Sat Nov 22, 2008 12:25 pm

What apache modules are you using? We've found that mod_memcache definitely causes this to happen on all systems its used on.

faris · Unread post by **faris** » Sat Nov 22, 2008 1:23 pm

removing mod_memcache doesn't seem to have any effect with respect to the segfaults on VPSes.

And as I recall it didn't have any effect on hostinguy's dedicated server either.

Unread post by **mikeshinn** » Sun Nov 23, 2008 12:27 am

And does this only happen with VPS systems?

Atomicorp

Apache segmentation fault

For those that see the segfaults

Any other modules installed?