tonight was the second time i've watching apache go off the charts and overload my server. we're talking about a server that rest of the time sits around with load averages of like .5 to 5 max shoot up to 80+

and both times looking at running processes it's httpd threads... frist time was 3 days ago and at first I was thinking DDOS or a site was slashdoted but a restart of apache and all would go back to normal for a while.

it's just throwing me for a loop why the thing is going so crazy

Is there anything useful in your apache logs? Do you have the server status module compiled in or enabled? If you catch it while its happening, you might be able to "see" whats going on with server status.

Frank

well just first impressions reading though i notice that there are a bunch of mysql access errors (max connections limit) then it starts giveing errors about not being able to kill child processes

bascially

1) 21:21 mysql errors start
2) 21:26 child process errors
3) 21:30 reset apache
4) 21:40 child process errors

and then basically repeats 3 and 4 and thats what's going on right now. reset and it's fine for a bit then boom it jumps back up...

keep top running and see what it could be. On one of my servers, on of my clients greymatter scripts were totally taking the server over, running the load up. The script would apparently hang and take up like 50 of cpu and mem and then they would load it one more time. It was ugly.

Do you use a control panel?

Do you have any max child process set in your httpd.conf? Check all your max's in that file and limit them?

we use webmin...

far as it goes, partner played with some apache settings and so far things have seemed to have gotten a bit more stable. also he setup a cron job to restart apache once an hour just to be safe...

i've been watching and checking logs and nothing really does it... no cgi, modules, etc. the child processes just decided it be real fun to go belly up and not stop...