Apache Problem...

[Anon]

I don't know this is the appropriate forum to ask about my apache problem but here we go.

I have a server running Apache 1.3.2 on RedHat Linux. 800Mhz 512MB RAM.

I get about 100,000+ hits a day. I'm looking at the server resources and there are several processes that are taking up 10:30 minutes of CPU time!!!! and about 5% of CPU. Also the threads seems to not die and regererate. This slows down my server soo much although bandwidth is low the server's httpd proceses are taking up 80+% of server resource.

About 239 proccesses are running at any time

MY httpd.conf is:

Timeout 150
Keepalive off
Min start server 5
Max start server 10
Start server 15
Maxchildspaw 50000
Max clients 256 (hard coded - no way to recompile apache due to Dell's messed up install)

Please advise on what I can do/

(No I can't get another server with another OS 😉)

[Anon]

Quick fix is to put a batch file in cron.hourly that does this:

killall -USR1 httpd
sleep 5
httpd

That will stop all apache children and gracefully restart them when they finish their current transfer. Should something go wrong, the sleep5/httpd will restart webserver, if it didn't stay up from the killall -USR1.

[Anon]

hmm.. i tried that and as soon as server starts all resources go up.. System is at 78%. THere are HTTPd threads that last 1:50 min... i've seen some that last 30 min!!!! The timeout is 120 and keepalive is off.

THere are some threaded that even take 16.7% of CPU time.

I don't know what to do. I stopped the server last night and then waited 2 min and started it again for the same number of proccesses it's having now, it was running at only 9.8% System CPU and everything was perfect.... but now it's back up to the high numbers

[smarlowe]

What all are you doing? I.e. which external modules are you calling, are you using persistant db connections, and is there any chance of upgrading your apache/php combo to the latest?

I thought Dell just installed RedHat.