[ntp:questions] ntpq and local clock driver

Eike Middell eike at middell.net
Mon Mar 23 05:25:52 UTC 2009

The main reason for my suspicion is a case of old faulty
GPS hardware which reported last years leap second twice.
This led to the question how to detect such errors.

But also in general it would be interesting to see how the
hosts controlled by ntp behave with respect to an independent
clock. If this is a normal PC clock one will most probably
see only its drifting. The controlled hosts should all
be stable. But if the drifting is continous, then serious 
errors could be seen from such a measurement. 

Thanks for the answer.
Eike Middell

On Mon, Mar 23, 2009 at 12:11:49AM +0000, Unruh wrote:
> emiddell at gmx.net (Eike Middell) writes:
> >Hi, 
> >I've got a question regarding a ntp daemon that has only the
> >"undisciplined clock driver" as a reference clock. To what 
> >extent does this daemon influence the local clock?
> Not at all. The local clock means that ntp does nothing. 
> >The context of the question is the following.
> >I've got a small network time synchronized to one host with a GPS clock.
> >The status of the individual ntp daemons (especially the offset to their
> >respective reference clock) are periodically monitored with ntpq.
> >I would like to compare this setup to an independent time reference in
> >order to see if there are collectively performed jumps. These I wouldn't
> >see in the monitoring that I have so far. Since I want to detect
> >suddenly occuring offsets, the independent reference would not have to
> >be frequency stable. It only has to provide a continous time line.
> Why are you suspecting jumps? 
> >A cheap solution would be to add another host to the network which
> >distributes its local time by ntp. So it would be possible to monitor
> >this independent source with the already available tools. Of course
> >nothing would be gained if the ntp daemon running on this host would
> >create jumps in the provided time.
> >Thanks in advance,
> >Eike Middell
> _______________________________________________
> questions mailing list
> questions at lists.ntp.org
> https://lists.ntp.org/mailman/listinfo/questions

More information about the questions mailing list