Apache memory madness

I have this problem on my heavily-loaded dedicated server (managed with WHM), where a single httpd process will spontaneously explode in size, consuming all the available RAM and bringing the server to its knees.

The process will get several hundered megs big in about a minute, driving load into the 90s.

It will only happen once every few days, and I have ulimit -v 12000 set in /etc/init.d/httpd, which kills the runaway processes before they become a problem.

But if /etc/init.d/httpd is reset (like it was just now somehow), I'm in trouble. It's extremely annoying. I hate it.

Is this is known problem with Apache or something else on my system? All my pages are in PHP, but with no potential for leaks. Turck MMCache installed, but the problem existed before iirc.

Any help would be greatly appreciated!

 

 

 

 

Top