We are in the process of migrating this forum. A new space will be available soon. We are sorry for the inconvenience.

OCO bringing my RPS to a treacle like crawl


Mikeh
22-10-2009, 11:57
Thanks, ive removed the three cron jobs to see if that makes any difference

Myatu
20-10-2009, 11:58
This is OCO: http://help.ovh.co.uk/Oco

Looking at the install script, it puts 3 OCO scripts (60secs, 120secs and 300secs) in the crontab /etc/cron.d/oco. These scripts use a locking file in the /usr/local/oco/lock/ directory. So my take is that if you don't use load balancing, then OCO is not needed (and thus, you should be able to safely remove it from the crontab).

Mikeh
20-10-2009, 09:19
No one? hmm ... the OVH engineer fixed this before but it has come back and they wont tell me the fix, so again I have to wait 24 hours for them ... sigh

Mikeh
01-10-2009, 11:28
Hi,

Need some urgent help on this one, basically my rig has been crawling along for the past 5-6 days, seriously slow, treacle like responses (e.g. 4 minutes to load WHM, 2 minutes to get the login on SSH), its only had 92mb memory free and been swallowing swap like nothing ive seen before (was using all 1gb of mem and 500mb swap).

A bit of tinkering, moving files around etc I noticed on the TOP command that I always had approx 8500+ process, of which 99% of them where sleeping and those which where sleeping all belonged to OCO. Done a bit of digging and run this command to kill OCO process's

kill -9 $(ps auxw | grep oco | awk '{print $2}')

And wow it was like having a brand new lightning speed server again! however within seconds of running the command the OCO processes kept coming back and the servers speed dropped back down to a crawl once again. I done a service oco stop and whilst this claimed to stop the server it kept spawing new processes, could this be on a cron?

So my question is ... what is the OCO exactly? (I know its an OVH monitor service?) can I just remove the thing from my box? obviously no point keeping a monitoring service running when the rig isnt evening running in a usable manor ...

Help


Thanks
Mike


EDIT (TOP as the server is currently, about 40 mins after running the last KILL command, the sleeping process increases until it hits 8000 odd)

top - 13:47:18 up 12:42, 3 users, load average: 8.91, 8.61, 7.77
Tasks: 5412 total, 2 running, 5410 sleeping, 0 stopped, 0 zombie
Cpu(s): 13.6%us, 15.2%sy, 0.0%ni, 18.8%id, 50.8%wa, 0.3%hi, 1.4%si, 0.0%st
Mem: 1022672k total, 770896k used, 251776k free, 3032k buffers
Swap: 1952692k total, 493460k used, 1459232k free, 12728k cached

PID USER PR NI VIRT RES SHR S %CPU %MEM TIME+ COMMAND
9765 root 20 0 5328 4108 784 R 15.1 0.4 0:03.78 top
10407 root 20 0 2172 820 708 R 10.1 0.1 0:00.36 ps
19784 squid 20 0 77596 51m 2788 D 7.3 5.2 34:25.21 squid
235 root 15 -5 0 0 0 D 2.0 0.0 3:22.19 kswapd0
3638 named 20 0 84328 27m 1288 S 2.0 2.8 6:10.17 named
10366 root 20 0 4240 2256 1536 S 1.7 0.2 0:00.06 check-ro-partit
10369 root 20 0 4240 2260 1536 S 1.7 0.2 0:00.06 check-ro-partit
1 root 20 0 2060 560 528 S 0.3 0.1 1:28.31 init
8 root 15 -5 0 0 0 S 0.3 0.0 1:02.26 ksoftirqd/2
1062 root 15 -5 0 0 0 S 0.3 0.0 1:08.88 scsi_wq_2


Now I run the kill command as above


top - 13:47:53 up 12:42, 3 users, load average: 36.60, 14.36, 9.66
Tasks: 222 total, 2 running, 220 sleeping, 0 stopped, 0 zombie
Cpu(s): 26.6%us, 28.4%sy, 0.1%ni, 22.9%id, 21.1%wa, 0.2%hi, 0.8%si, 0.0%st
Mem: 1022672k total, 253252k used, 769420k free, 3508k buffers
Swap: 1952692k total, 144684k used, 1808008k free, 14168k cached

PID USER PR NI VIRT RES SHR S %CPU %MEM TIME+ COMMAND
11251 root 20 0 4456 2272 1436 R 7.7 0.2 0:00.04 oco-tcpresponde
1 root 20 0 2060 560 528 S 0.0 0.1 1:29.04 init
2 root 15 -5 0 0 0 S 0.0 0.0 0:00.00 kthreadd
3 root RT -5 0 0 0 S 0.0 0.0 0:01.92 migration/0
4 root 15 -5 0 0 0 S 0.0 0.0 0:59.64 ksoftirqd/0
5 root RT -5 0 0 0 S 0.0 0.0 0:02.12 migration/1
6 root 15 -5 0 0 0 S 0.0 0.0 1:04.83 ksoftirqd/1
7 root RT -5 0 0 0 S 0.0 0.0 0:01.98 migration/2
8 root 15 -5 0 0 0 S 0.0 0.0 1:02.29 ksoftirqd/2
9 root RT -5 0 0 0 S 0.0 0.0 0:02.08 migration/3

Now the loads do come down, but the load only seems to be on there because OCO seems to want to respawn continuously making all these processes and consuming my memory.