I have several machines running 16.10, both servers and laptops, all on the same network. All of them occasionally have the same problem with NTP. The clock drifts slightly out of sync and when I check the NTP peers there is only one and it's not being used.
> ntpstat
unsynchronised
polling server every 8 s
> sudo ntpq -p
remote refid st t when poll reach delay offset jitter
==============================================================================
10.x.x.x .INIT. 16 u - 64 0 0.000 0.000 0.000
I've tried with the default Ubuntu ntp.conf
server list (from 0 to 3) of
pool 0.ubuntu.pool.ntp.org iburst
...
And also with the standard ntp list (also from 0 to 3) like so:
server 0.pool.ntp.org
...
But the issue is the same. The IP address of the unused server is always the same on all the machines, regardless of their ntp.conf
.
I can temporarily solve it by making a change to ntp.conf
and restarting ntpd
but would like to fix the source of the problem so that useful peers are found automatically.
What can I try to fix this?