0
Suddenly one of my Linux VM is two minutes behind than the ntp server. I restarted NTP and it got fixed. I should have run ntpq -p before restarting. :-/
10 days ago it became 2 minutes behind, before it that was 1 minute behind and I'm not sure when it started. I only noticed it today and could track up to 10 days.
I'm curious why it can go behind? Could it be due the small ntp delays? Which means, it keeps going back few milliseconds once a while and then it's suddenly 2 minutes behind.
If that is the case, why would a ntp restart pick up accurate time when the poll couldn't have fixed it long ago?
I was not sure if the VM was polling ntp server, So I enabled logging and looking at the loopstats and peerstats it is polling.
This output is after I restarted ntp.
remote refid st t when poll reach delay offset jitter
==============================================================================
*x.x.x.x x.x.x.x 4 u 22 64 377 0.618 -1.980 0.768
So the output you provided, is the output when the time was off by 2 minutes, or when the time was correct? Because the offset I believe is in ms which means you are barely 3/4 of a second behind your NTP server which is pretty normal. "I'm curious why it can go behind?" - A lose of a connection to the NTP would cause the machine to start to do it's own thing. 'Could it be due the small ntp delays?" - Yes, but your values are not large enough to cause the behavior you describe – Ramhound – 2017-06-19T18:00:52.043
It was after the time was corrected – anon – 2017-06-19T18:01:29.047
Which means the output isn't really helpful. The output you provided is normal. – Ramhound – 2017-06-19T18:02:20.243