ntpd instead of systemd-timesyncd

[Rares]

Hi,

After an "apt upgrade" and reboot I got in cehp "cluster [WRN] 1 clock skew 0.131975s > max 0.05s". Indeed after 2 minutes the error dissapear from ceph dashboard.

Reading more I found:

systemd-timesyncd does no clock discipline: the clock is not trained or compensated, and internal clock drift over time is not reduced. It has rudimentary logic to adjust poll interval but without disciplining the host will end up with uneven time forever as systemd-timesyncd pushes or pulls at whatever interval it thinks the near-term drift requires. It also can't assess the quality of the remote time source. You're unlikely to get accuracy much greater than 100ms. This is sufficient for simple end user devices like laptops, but it could definitely cause problems for distributed systems that want greater time precision.

And systemd-timesyncd seems to do the corrections with time steps and not by skewing time.

Would it be an idea to replace systemd-timesyncd with ntpd in proxmox?

Regards,
Rares

 

[wolfgang]

Hi,

no there are no plans. But you are free to use what you like. NTPD is also not perfect and has some problems.

 

[Rares]

Thank you Wolfgang for the reply.

My question was if there are any benefits in using ntpd instead of systemd-timesyncd? Also do you think that having a local ntp server would improve the stability of the cluster? Yould you invest in a local GPS controlled dedicated NTP server device?

Regards,
Rares

 

[wolfgang]

What is the exact usecase?
If just the cluster (corosync) there is no need for it.
But as VM has always a time drift you should have a local NTP server and use this.
Also, it depends on the application if you need the correct time or just an identically time?