I have a 768 Slice at slicehost.com running Ubuntu Server 8.04.2 LTS (hardy) with a LAMP stack on it that periodically crashes, though why I am not sure.
From what I can tell, there is a process that basically goes rogue and consumes all the memory on the slice, suffocating all the other programs running until the whole thing comes to a grinding halt, and I have to do a hard reboot of the slice to get it back up and running again. I can't detect any pattern for this (it seems to happen about once a month, more or less).
Here's a screenshot of my console during the last crash:
I would assume that a possible cause might a PHP script or an apache configuration rule that might cause the crash if triggered? How would I be able to find out which one is the offending one? I've checked and rechecked all my PHP scripts, and running them doesn't seem to trigger the crash.
I've also been able to log on to my system during a crash and see what's running (with top), but I can't tell how the offending process was started, so I can't trace the root of the problem!
I know my description is overly generic, but unfortunately my expertise in tracking down the source of these glitches is very limited. If you need any additional information about my system in order to help me figure this out, please let me know in the comments, and I will append it to the question.
My only other lead as to the culprit here is Wordpress, which we have installed on this server. Here are the details:
- Wordpress 3.0.3 with the following plugins installed and activated:
Addmarx - Bookmark/Share/Email Dropdown,Akismet,All in One SEO Pack,Animated Banners,Automatically publish highlights of any website, directly to your Blog,Broken Link Checker,CMS Dashboard,Collapsing Categories,Status Updater,SubHeading,Ultimate Google Analytics,VastSubCat,WP-CMS Post Control, andWP Super Cache
Result of free (not during crash):
total used free shared buffers cached
Mem: 765172 456044 309128 0 34748 243532
-/+ buffers/cache: 177764 587408
Swap: 1572860 0 1572860
freeoutput, though this is under normal load. Since I can't predict or manually trigger a crash, I can't sent the output during a crash. – neezer Dec 28 '10 at 19:33Apache (internal dummy connection)entries, which all look identical...just different timestamps. I have numerous (over 10) sites vhosted on this server, each with their own access log. I'll take a look at them to see if I can find anything anamolous... what should I be keeping an eye out for? I guess if I can't find anything there I'll have to post them all individually... – neezer Dec 28 '10 at 19:37