IPB started to act weird....Please help.

[awaisuk]

IPB started to act weird....Please help.

Well first of all I have not touched my server in awhile now, did not install any new mods, and for the past 2 days my site has been really slow.

CPU Load: 12+

Recount and rebuild of forums are also not working, it is not updaing the last poster on the forums.

top - 14:50:58 up 23:06, 1 user, load average: 15.90, 16.68, 14.81
Tasks: 176 total, 23 running, 152 sleeping, 0 stopped, 1 zombie
Cpu(s): 87.2%us, 11.8%sy, 0.0%ni, 0.0%id, 0.0%wa, 0.2%hi, 0.8%si, 0.0%st
Mem: 2067432k total, 2013704k used, 53728k free, 6016k buffers
Swap: 4096564k total, 112k used, 4096452k free, 1634452k cached

PID USER PR NI VIRT RES SHR S %CPU %MEM TIME+ COMMAND
29399 apache 15 0 35512 20m 3428 S 17.2 1.0 0:12.38 httpd
29831 apache 16 0 34156 18m 3360 R 13.6 0.9 0:07.55 httpd
23973 apache 15 0 34948 20m 3500 R 11.6 1.0 0:45.02 httpd
30010 apache 16 0 36216 20m 3436 S 11.6 1.0 0:07.55 httpd
30379 apache 15 0 32596 17m 3428 S 11.2 0.9 0:04.26 httpd
25959 apache 15 0 32628 18m 3832 S 10.6 0.9 0:29.20 httpd
30616 apache 16 0 28496 13m 3296 S 10.2 0.7 0:03.30 httpd
30646 apache 15 0 31344 16m 3348 S 10.2 0.8 0:02.90 httpd
30630 apache 15 0 31060 16m 3296 S 9.6 0.8 0:02.53 httpd
29258 apache 15 0 30284 15m 3432 S 8.9 0.8 0:05.16 httpd
30645 apache 16 0 33604 18m 3400 S 8.6 0.9 0:03.30 httpd
29236 apache 15 0 33136 17m 3416 R 7.9 0.9 0:08.99 httpd
29491 apache 15 0 32340 17m 3416 S 7.6 0.9 0:11.28 httpd
30648 apache 16 0 32720 17m 3344 S 7.6 0.9 0:02.49 httpd
24744 apache 15 0 37924 13m 2536 R 5.0 0.7 0:10.96 python
28944 apache 15 0 31344 16m 3432 S 4.6 0.8 0:12.67 httpd
30008 apache 15 0 31108 16m 3428 S 4.3 0.8 0:05.79 httpd

Please help me fix this.

Thanks.

[3DKiwi]

A CPU load of 12 is serious and mine only goes over 1 on a dual core dedicated server only at peak times and usually for short periods. If at all possible get your server rebooted. if problems persist talk to your host and then if that doesn't help talk to IPS.

3DKiwi

[Will Munny]

awaisuk, you're not the first to experience this problem and it has yet to be identified properly. There is an issue with the software in certain environments that has yet to be identified and examined properly. There are a couple of threads running here, with admins experiencing the same symptoms as you. I was experiencing loads of over 200, randomly and for no apparent reason.

Are you using eAccelerator? If so, it appears to be related. I managed to stabilise my server by removing eAccelerator and installing Xcache as an alternative.

[awaisuk]

I am not using anything extra, I just installed IPB on there and it was working perfectly fine, and running fast, but then it automatically got slow, and even a reboot does not help.

[dyelton]

awaisuk, you're not the first to experience this problem and it has yet to be identified properly. There is an issue with the software in certain environments that has yet to be identified and examined properly. There are a couple of threads running here, with admins experiencing the same symptoms as you. I was experiencing loads of over 200, randomly and for no apparent reason.

Are you using eAccelerator? If so, it appears to be related. I managed to stabilise my server by removing eAccelerator and installing Xcache as an alternative.

I'm having the same issue and it started happening the day I ran easyApache from WHM/cPanel to upgrade to PHP 5.2.12 (I have since upgraded to 5.3.1 and the issue still persists). I don't attribute it to PHP however, easyApache only lists eAccelerator v0.9.6 RC1 now instead of the older non-RC version that I previously had with PHP 5.2.11.

The high load seems random and every time it happens it goes up to around 260 on my server...which obviously brings it to its knees. I may go a day without an issue, or I may only make it an hour. My host is blaming it on IP.Board, but I'm pretty certain that isn't the root cause. I do have a busy forum, but I ran IPB 3.0x since the end of October for two months straight without any issues to speak of.

Anyone else having this issue and can correlate it with the 0.9.6 RC1 version of eAccelerator?

[dyelton]

Looks like there may be some meat to this...check out this bug report for eAccelerator v0.9.6 RC1:

http://eaccelerator.net/ticket/368

[Will Munny]

All I can tell you is that I was never able to diagnose the cause of the overloads... but since dumping eAccelerator, I've not had any problems. Xcache is functioning very well in my environment (touch wood).

[dyelton]

After posting in this thread I disabled eAccelerator and installed APC. Last night was the first night in a week or so that I've not been awakened by a text message letting me know that my server had high load. :)

I chose not to go with xcache over some reports that it doesn't play well with PHP 5.3.1. APC is SUPER easy to install...should have done this long ago and not messed around with eAccelerator.