On 2/27/2012 10:15 AM, Donny Brooks wrote:
Here recently we have been having a major problem with high load on
our SOGo server. It will go from a normal load to unresponsive in less
than 5 minutes. If I don't catch it in time and stop/start the
services it will require I power the server off and restart it. The
last lines in the sogo.log before it hangs is this:
Feb 27 10:05:50 sogod [1877]: [WARN] <0x0xf8fc7a8[WOWatchDogChild]>
pid 3184 has been hanging in the same request for 1 minutes
Feb 27 10:05:54 sogod [1877]: [WARN] <0x0xf8db168[WOWatchDogChild]>
pid 3185 has been hanging in the same request for 1 minutes
Feb 27 10:05:54 sogod [1877]: [WARN] <0x0xf8fbae8[WOWatchDogChild]>
pid 3217 has been hanging in the same request for 1 minutes
Feb 27 10:05:54 sogod [1877]: [WARN] <0x0xf8f01e8[WOWatchDogChild]>
pid 1994 has been hanging in the same request for 4 minutes
Feb 27 10:05:55 sogod [1877]: [WARN] <0x0xf926298[WOWatchDogChild]>
pid 3054 has been hanging in the same request for 1 minutes
I have set the worker count to 8 (prefork setting) and the SvMemLimit
to 256. The machine has 6 cpu cores at 2.0Ghz each and I have given it
6GB ram. I am at a loss. Any input would be great. Just really looks
bad that we did not have these issues until now, just when we roll it
out in production...figures huh.
Donny B.
Sorry forgot to say this is a Xen pv domu running Centos 5.7. Nothing
wrong with any of the other virtual machines except this one.
--
[email protected]
https://inverse.ca/sogo/lists
