NOHZ: local_softirq_pending 100 - is there something to worry about? [message #45131] |
Tue, 31 January 2012 20:07 |
insider
Messages: 11 Registered: January 2012
|
Junior Member |
|
|
Hello,
Now again, after the node reboot because of network device freeze (e1000e driver bug? Still not updated driver version in openvz kernel) our node up for 22hours and we see this message in the logs:
NOHZ: local_softirq_pending 100
Is this is something we have to worry about? For a while it is not caused any problems. Or we should just ignore this message.
P.S.
Well, I think we made a mistake to put RHEL6 openvz 2.6.32 kernel on the production to early... the more we use it the more bugs, locks, freezes and problems we have. It is far far away from "stable"...
All other 2.6.18 centos 5 openvz nodes runs very good.
Considering now to go back to RHEL 5 with 2.6.18 kernel for one more year, even if End-of-life for RHEL5 is alot shorter than RHEL6...
Or maybe there is solution to use 2.6.18 kernel in Centos 6 and not to use 2.6.32 until it becomes really "stable" and then just switch to a 2.6.32? What you think about it?
Thank you for any answers of suggestions.
|
|
|