Oh that is interesting ...
# ntpctl -s all
0/1 sensors valid, clock unsynced
sensor
wt gd st next poll offset correction
vmmci0
1 0 0 13s 15s - sensor not valid -
sensor not valid - why would that happen?
I have no idea about the inner workings of vmd .. but maybe this host
has
so many VMs that it is hitting some limit?
Don't forget that ntpd requires numerous polling cycles (at 15 seconds
each)
to convince itself about how reliable it's time sources are.
Give it 5 minutes after restart, or even longer, before deciding for
yourself
that it is not going to sync.
Another reason it may not sync is that the client current time is "too
far away" from
the external ntp time sources. You need to jump the time manually to
fix that.
--John
