Hello, Im having some problems.. i have a top ofthe range dual server, with two sites running on it with good traffic.. however, they are both going VERY slow when i get traffic surges.. the load stays at around 0.7-1.... however, i keep getting processes that say defunct or something, and that's what apparently is causing the problems.... below is a copy of my "top" - PLEASE someone give me simple advice on what to do, im a real newbie!

3:02pm up 1 day, 10:39, 3 users, load average: 0.31, 0.67, 0.62
156 processes: 142 sleeping, 1 running, 13 zombie, 0 stopped
CPU0 states: 4.3% user, 7.2% system, 0.0% nice, 87.3% idle
CPU1 states: 8.1% user, 6.3% system, 0.0% nice, 84.4% idle
CPU2 states: 9.1% user, 6.4% system, 0.0% nice, 83.3% idle
CPU3 states: 6.3% user, 5.0% system, 0.0% nice, 88.0% idle
Mem: 1030560K av, 747604K used, 282956K free, 0K shrd, 82304K buff
Swap: 2048276K av, 0K used, 2048276K free 506560K cached

PID USER PRI NI SIZE RSS SHARE STAT %CPU %MEM TIME COMMAND
15916 nobody 16 0 6040 6040 4144 S 6.7 0.5 0:00 httpd
15937 nobody 16 0 0 0 0 Z 5.5 0.0 0:00 httpd
12654 root 16 0 3860 3860 3720 S 4.3 0.3 0:46 httpd
15927 nobody 17 0 5692 5692 4136 S 4.1 0.5 0:00 httpd
1711 mysql 15 0 21768 21M 2000 S 0.7 2.1 0:03 mysqld
13754 root 15 0 1112 1112 848 S 0.7 0.1 0:00 top
15630 root 15 0 1112 1112 848 R 0.7 0.1 0:00 top
27588 mysql 15 0 21768 21M 2000 S 0.5 2.1 0:04 mysqld
28497 mysql 15 0 21768 21M 2000 S 0.5 2.1 0:04 mysqld
32015 mysql 15 0 21768 21M 2000 S 0.5 2.1 0:03 mysqld
15943 nobody 16 0 0 0 0 Z 0.3 0.0 0:00 httpd
15945 nobody 16 0 0 0 0 Z 0.3 0.0 0:00 httpd
15949 nobody 16 0 0 0 0 Z 0.3 0.0 0:00 httpd
1 root 15 0 488 488 420 S 0.0 0.0 0:07 init

Restart apache. Stop it then wait about 15seconds then start it again. Dont just restart it. This is a major pain in the hiney with Cpanel. Apache processes loop out of control and they cant or will not fix it! Amazing!

Hi Aussie,

Tell me something. How is it that Apache would cause Exim to have so many processes? Here is one of my recent top's

10:17pm up 2:46, 2 users, load average: 34.33, 32.36, 24.35
794 processes: 283 sleeping, 2 running, 508 zombie, 1 stopped
CPU states: 11.2% user, 44.8% system, 0.0% nice, 43.9% idle
Mem: 506048K av, 500572K used, 5476K free, 0K shrd, 214320K buff
Swap: 2048276K av, 80012K used, 1968264K free 98368K cached

PID USER PRI NI SIZE RSS SHARE STAT %CPU %MEM TIME COMMAND
23119 root 10 0 2836 2836 2220 S 14.9 0.5 0:00 exim
23135 root 18 0 2800 2800 2220 S 9.3 0.5 0:00 exim
15429 root 11 0 1024 1024 396 R 6.5 0.2 0:49 top
23131 mailnull 14 0 3004 3004 2360 D 6.5 0.5 0:00 exim
20538 mailnull 10 0 1596 1436 952 D 5.6 0.2 0:10 exim
21500 mailnull 13 0 1556 1556 952 D 5.6 0.3 0:06 exim
21499 mailnull 10 0 1668 1612 1060 D 4.6 0.3 0:05 exim
23126 root 13 0 2828 2828 2220 S 4.6 0.5 0:00 exim
23133 root 14 0 2804 2804 2220 S 4.6 0.5 0:00 exim
21505 mailnull 12 0 1556 1556 952 D 3.7 0.3 0:06 exim
21557 mailnull 10 0 1560 1520 952 D 3.7 0.3 0:05 exim
23113 root 13 0 2824 2824 2220 S 3.7 0.5 0:00 exim
23121 mailnull 9 0 3456 3456 2748 S 2.8 0.6 0:00 exim
23127 root 12 0 2704 2704 2124 S 2.8 0.5 0:00 exim
15414 nobody 10 0 4580 4040 580 S 1.8 0.7 0:03 httpd
20529 mailnull 12 0 1588 1568 952 D 1.8 0.3 0:11 exim
20677 mailnull 10 0 1576 1576 952 D 1.8 0.3 0:09 exim
23012 mailnull 9 0 2140 1760 1556 D 1.8 0.3 0:00 exim
23120 mailnull 9 0 2868 2868 2256 S 0.9 0.5 0:00 exim
23132 mailnull 11 0 2896 2896 2256 D 0.9 0.5 0:00 exim
23137 root 12 0 1492 1488 1208 D 0.9 0.2 0:00 exim
1 root 6 0 112 64 48 S 0.0 0.0 0:06 init
2 root 8 0 0 0 0 SW 0.0 0.0 0:00 keventd
3 root 9 0 0 0 0 SW 0.0 0.0 0:00 kapmd
4 root 18 19 0 0 0 SWN 0.0 0.0 0:00 ksoftirqd_CPU0
5 root 9 0 0 0 0 SW 0.0 0.0 0:00 kswapd
6 root 9 0 0 0 0 SW 0.0 0.0 0:00 kscand/DMA

If I restart Apache (httpd) almost all of the exim processes (showing in Zombies) goes away. OR of I restart Exim they will go away. I don't get it and am going bald trying to work this out. I have to set 2 crons to activate each half hour to eliminate the zombies or my server starts to crawl. 1 CRON to restart httpd and 1 CRON to restart Exim. But in doing this, I also find that my maiil que gets slam full. Just today I cleared our over 12gigs of old email that I didn't know was there until 2 days ago.

I have the current versions of CPanel, Exim, Kernel

Any ideas? Anybody?

Thanks
Michael

Maybe somebody is using your apache to send out spam via some web based form or something?

I thought about that but have monitored the server closely and find that that is not so. Netstat and PS are telling me its exim processes that are stuck. As if they aren't getting a terminate signal or something. I dunno. Its making me crazy and putting excessive wear and tear on my server. I have 2 servers with the same hardware. 1 with Ensim (Which has been little to no trouble at all) and CPanel which I'm beginning to not like at all. Its way to much work to maintain.

Thanks for the reply though!

Happy New Year.
Michael

I've heard that if you're using an edge release, that updating to the latest one as of yesterday seems to resolve the zombie bug.

Chew

Hey Chew,

Forgive my ignorance here but, an edge release of what?

I think he means the Cpanel version after updating. You can use Edge, Release and stable versions.

Thanks Orca.

I thought so, just wanted to make sure.

Thanks again
Michael

I like to stay with stable because for the most part it keeps working. Release is the default value for the updating and it can cause troubles at times. Of course stable has trouble from time to time...

I agree. But, I think I had CP setup for the release version instead of the stable (Preferred) sooooooooooo, I'm changing that now. I hope this eliminates this problem. I updated the kernel the other night and my load averages got even worse. If this problem continues I'm requesting this box be changed to something else. I don't mind managing my boxes, I just don't like having to deal with the same issues since the first day I leased this box. RS has always been on the ball with support issues for us but I think we are dropping the ball on this particular box. We'll see.


Thanks for the info though

Michael

Yes it is setup for release by default, thats cpanel's doing. As far as being bad from the first day, that could have been a bad image or something. In the future you might try and goto cpanels site and reinstall cpanel, its only one command, then run the common fix scripts. To date I have been able to eliminate problems on a FRESH box by doing this.