Hi Everyone,

Did anyone encountered leap second issue in rhel 6.3?
We recently had a production outage due to this issue which occurred on 1st Jan 2017 00:00 AM.
It literally spoiled my new-year party and i don't want to encounter this again next year.
Due to product obligations, we can't update the OS and hence need to find out a workaround.

Red-hat documentations says to run ntp in skew mode, i just wanted to know from the community that did anyone encountered this issue and if yes, what workarounds actually worked apart form upgrading the OS?

Thanks,
Mani Prakash

No.

After the RedHat announcement about this past leap second before it occurred I had done reading that suggested we really only have to worry if we do an absolute time update at exactly midinight UTC when the leap second occurs. Since we had a leap second in 2015 without any noticeable ill effects it seemed likely we wouldn't this time either.

We in fact saw no issues on any RHEL5/RHEL6/RHEL7 server due to this leap second.

If you saw an issue you may want to look into what might be doing an absolute time update at exactly midnight UTC on your system.

I don't think there is much benefit to addressing it now, after the fact. You may need to address it if they announce another coming leap second in future.

1 members found this post helpful.

rhl 6.4 and above has this issue fixed, we encountered it specifically on rhel 6.3 servers.
This resulted in high CPU utilisation caused because of the java processes and we had to reboot the vms.

We have a mix of releases including 4 that are on RHEL6.3 and others that are on RHEL6.0 through RHEL6.2. None of them experienced the issue.

I've seen it back on 2015 on a Centos-6.2 running zimbra (that is a java based mailserver). It's actually a java problem, so updating just java (not the whole OS) will solve it.

Anyway at that time there was no java without the leap second issue available, so I had to run the following commands (from here):
Then restarting the problematic application (zimbra), everything turned back to normal

HTH

The most simple command to clear the leap second flag is
I would simply upgrade the OS to 6.9 or at least 6.4
If this is not an option you can install a 6.4 kernel on 6.3

Hi,

The leap second problem was due to the java version shipped with RHEL-6.2/6.3.
So OP can just update java if upgrading OS is not an option, or else change the time and restart ntpd if needed

Regards

Leap second has become a pretty common problem and some of the common issues can be observed during this time are:
- High CPU usage for processes like Java
- Sometimes even VM becomes unresponsive

Resolution:
- To resolve this either we can go ahead with Patching process and if we don't want to go with that then there is another way also.
- Manually resetting the date using below command can also fix this issue.
date -s "`LC_ALL=C date`"
OR
date `date +'%m%d%H%M%C%y.%S'`

More details can be found here (may ask for login)
https://access.redhat.com/solutions/154793

get the below link to resolve the issue
https://access.redhat.com/labs/leapsecond/
https://access.redhat.com/solutions/1471933
https://access.redhat.com/solutions/1530053