Leap second 2016

Latest response

As you guys probably know, on December 31 a leap second will occur.
I am well aware of the need to update the kernel, NTP and tzdata packages.
My question is: if my servers were prepared for the 2015 leap second, meaning I updated the kernel, tzdata and ntp packages to my servers so they won't be affected in 2015, do I need to update the packages again to overcome this year's leap second? Or are the 2015 "safe" kernel, tzdata and ntp packages still safe for this leap second?

Responses

You can find all the details in this article.

I have read the entire article. Non of it talks about last year's leap second and whether the last update was enough,,.

You have to upgrade the tzdata package to a version higher than tzdata-2016f-1 - the latest version as of now is : tzdata-j-1.
Having the latest available stable kernel installed is more than recommended. Please check whether the chrony daemon is running : systemctl status chronyd - in case that the daemon is not running (it should be enabled by default), execute these commands :

sudo setenforce 0
sudo systemctl enable chronyd
sudo setenforce 1

Why do you need to set SELinux to permissive mode before enabling chronyd?

Hi Jörg ! Because otherwise the command did not work ... in my case at least. :)

Well, that seems a little bit odd. I just checked on a test machine. There it worked without changing the SELinux mode.

Maybe it has got something to do with the fact that in this case RHEL is installed in a virtual machine - on a bare metal installation chronyd was already running by default.

I would guess that there is some other issue. I run RHEL 7 in virtual machines only. But maybe this is an issue for a support case.

Possible, but I think it also might have been the version of the SELinux policy module when I ran the virtual machine on my fedora 25 host system with KVM / virt-manager. I experienced something like this in other cases and once a new SELinux policy version was released, the same thing worked without any interactions.

Chrony is only recommended for systems that are regularly suspended, not servers that permanently on. Having chronyd running is not a prerequisite to overcoming the leap second challenges if you are running ntpd.

From the documentation:

Chrony should be considered for all systems which are frequently suspended or otherwise intermittently disconnected and reconnected to a network. Mobile and virtual systems for example.
The NTP daemon (ntpd) should be considered for systems which are normally kept permanently on.

Thank you very much for this information, now I know that there was no issue with my VM setup ! :)

Close

Welcome! Check out the Getting Started with Red Hat page for quick tours and guides for common tasks.