Serviceguard Error Mesage

[Christian Marquardt_1]

Hello,
we've the following message in syslog.log:

May 9 09:12:28 hpolpd1 cmcld: Timer_loop delayed: current state=1 pop=(0,379433111), now=(0,379433385), delta=2s(0,274)
May 9 09:10:01 hpolpd1 su: + tty?? root-pol292
May 9 09:12:28 hpolpd1 cmcld: Timer_loop's previous check_timers started at tsb (0,379433091) and lasted 0s (0,0) executed 3 callbacks
May 9 09:10:03 hpolpd1 su: + tty?? root-pol192
May 9 09:12:28 hpolpd1 above message repeats 3 times
May 9 09:12:28 hpolpd1 cmcld: Timer_loop's previous sigwait started at tsb (0,379433091) and lasted 2s (0,294)
May 9 09:12:28 hpolpd1 cmcld: Timer_loop's previous cm_lock started at tsb (0,379433385) and lasted 0s (0,0)
May 9 09:12:28 hpolpd1 cmcld: Timer_loop's last timer callback (type=35,id=0) started at tsb (0,379433091) and lasted 0s (0,0)
May 9 09:12:28 hpolpd1 cmcld: Timer_loop's last greater than 1s timer callback (type=0,id=0) started at tsb (0,0) and lasted 0s (0,0)


Can everyone explain me what they means??? Can't find any information via google.

regards
Christian

[Rita C Workman]

Hi Christian,

I have never seen that error message before. You don't mention your environment or anything about your SG version that might have helped.

I did find this one thread where somebody had the same error on their Linux cluster that might help. Very little info, and the submitter did not bother to add any final information or resolutions.....but it might help you:

http://search.hp.com/redirect.html?url=http%3A//forums1.itrc.hp.com/service/forums/questionanswer.do%3FthreadId%3D1074737&qt=%2Bcmcld%3A+%2BTimer_loop+&hit=1

Rgrds,
Rita

[Duncan Edmonstone]

Christian,

These messages could indicate that the cmcld process is unable to get any CPU time - this doesn't usually happen as cmcld runs with a real-time priority.

This then suggests that some sort of hang occurred at this time either due to other processes with RT priorities (there shouldn't be any!) or due to some kind of kernel/PDC issue.

More data would help - a bit more background about the environment and whats running on the boxes etc. It's always good to make sure you are bang up to date on Serviceguard pacthes as well...

HTH

Duncan

[Stephen Doud]

cmcld must run periodically to insure node and cluster viability. The messages you are seeing indicate it is not getting into the CPU in a timely manner. There are several historical causes for this and most likely causes that have not yet been documented.
I suggest you open a software case with HP to investigate the source of the bottleneck.