CUCM 7 - CallProcessingNodeCpuPegging

[trevor.mcnamara]

Hello,

I have alerts setup so that all errors, info etc.. get emailed to me.

Every day at midnight I get an email with the Subjet:[RTMT-ALERT-StandAloneCluster] CallProcessingNodeCpuPegging

And the following text:

Processor load over 90 Percent. cmoninit (62 percent) uses most of the  CPU. Processor_Info: For processor instance _Total: %CPU= 99, %User= 68,  %System= 16, %Nice= 0, %Idle= 0, %IOWait= 16, %softirq= 0, %irq= 0. For  processor instance 0: %CPU= 99, %User=  68, %System= 16, %Nice= 0, %Idle= 0, %IOWait= 16, %softirq= 0, %irq= 0.

So from what I can gather cmoninit is using a heap of CPU at this time everyday, however the system should barely be even used, we have an IVR however at this time of day, the system should be pretty much idle.

This is always generated by the Publisher.

It is probably not a problem, however I though I would find how why it is happening and also how to can be stopped.

I know I can turn the alert off however I dont want to turn it off in case i missed a actual legit issue.

Thanks

[Aaron Harrison]

Hi

cmoninit is an Informix process; I would guess that your backup runs at that time and that is the source of the high cpu usage?

Perhaps try moving the backup schedule an hour back or forward and see if the time of the alarm changes to verify this...

Aaron

Please rate helpful posts...

[trevor.mcnamara]

I actually dont have any backups scheduled at all...?

Is there a way I can find out exactly what is causing it to happen? Log files etc... I know if it was a normal Linux system, I would just check out cron etc... however I dont have the ability to do that.

[philippe.cousineau]

I have the same thing at midnight, and my backup are a t 02:00 am..

any idea ?

[DWAM_2]

Hello,

Have you found the solution ?

I've got a cluster of 5 ccm version 6.1.4.2190-3.

On one of this callmanager, I've got the same behaviour at 00H02 am every day :

Processor load over 90 Percent. cmoninit (6 percent) uses most of the CPU. Processor_Info: For processor instance 1: %CPU= 96, %User= 11, %System= 3, %Nice= 0, %Idle= 4, %IOWait= 82, %softirq= 1, %irq= 0. For processor instance _Total: %CPU= 97, %User= 11, %System= 3, %Nice= 0, %Idle= 3, %IOWait= 81, %softirq= 1, %irq= 0. For processor instance 0: %CPU= 97, %User= 12, %System= 3, %Nice= 0, %Idle= 3, %IOWait= 80, %softirq= 1, %irq= 0. The alert is generated on Thu Feb 23 00:01:49 CET 2012 on node.

This message appears recently.

I've checked the backup : it's done every day at 01H00 am.

Thank you.

[joshua.gertig]

Interesting post. We have noticed a lot of peculiar behavior on our system (cucm 8.5 vm on ucs chassis) at midnight. We see SDL link loss and phones re-register. In our case, this began after a shutdown of the primary, failed over to secondary UCS, the failed back to primary.

There is a DB maintenance window that occurs at midnight (be default) that may be the cause behind my issue.

If you've ruled out back-ups other things you may want to check is if LDAP re-syncs occur at that time, or CER Phone updates, other SNMP apps (like Singlewire), CDR delivery, QM applications that run updates, etc. Good luck!