After applying this year's Red Hat leap second patch, will the high CPU utilization issue from 2012 be triggered again this year?

[kscher]

Referring to an Answers item from 2012, We've got this year's Red Hat leap second patch applied. I do not know whether the high CPU condition described in the 2012 issue will be triggered again this year, but I'd like some guidance from Splunk and/or the community.

Anybody?

Thanks in advance...

[jrodman]

I haven't personally tested, but my understanding was that this behavior came from a Linux kernel problem, subsequently fixed. The same high cpu problem was reported in many other applications during that time interval, and could be quiesced by simply setting the system date to the date it already had.

So assuming the Linux kernel team has fully resolved the problem, and you've updated your kernels since 2012, it should not reoccur.

Personally I'm encouraging an internal test at Splunk Corporate for the leap second traversal now. We can't cover all cases, but we should cover the common one.

[dolivasoh]

Checking in, my cluster is still alive. Anyone know where this was officially patched?

[kscher]

Thanks Joshua,

> The same high cpu problem was reported in many other applications during that time interval, and could be quiesced by simply setting the system date to the date it already had.

Your answer goes some way toward helping me understand workaround specified in the 2012 item where you're supposed to stop ntp, execute date, and then restart ntp.

BTW, I've also opened an issue with Splunk to get the official line on this, and we'll see what they come up with.