[rhelv6-list] CPU utilization issue following leap second insertion
David Coulson
david at davidcoulson.net
Sun Jul 1 02:17:34 UTC 2012
Most of my RHEL6 systems running Java applications went nuts when the
leap second was added:
[root at rhtsutility02 stuff]# ssh rhesdevwesb01 grep UTC /var/log/messages
Jun 30 19:59:59 rhesdevwesb01 kernel: Clock: inserting leap second
23:59:60 UTC
Output of sar during interval in question
07:59:01 PM all 0.12 0.00 0.20 0.00 0.00 99.68
08:00:07 PM all 8.18 0.02 4.14 0.00 0.00 87.66
08:01:07 PM all 70.08 0.00 29.92 0.00 0.00 0.00
08:02:07 PM all 69.49 0.00 30.51 0.00 0.00 0.00
08:03:07 PM all 69.61 0.00 30.39 0.00 0.00 0.00
08:04:07 PM all 69.76 0.00 30.24 0.00 0.00 0.00
I've RHEL4 and RHEL5 systems running the same apps, same JVM, same J2EE
(mostly WebSphere), and they were all fine.
A couple of my RHEL6 boxes didn't log the 'inserting leap second' dmesg
entry, and they did not exhibit an issue - Not sure why they did not log
it, since they are all using the same NTP clock sources.
Anyone else see this? I'm going to grab a bunch of sosreports and get a
case open, but wanted to see if it was something unique to my environment.
More information about the rhelv6-list
mailing list