[ntp:questions] Analyzing Loopstats
Danny Mayer
mayer at gis.net
Tue Jul 12 13:08:53 UTC 2005
Jason C. Wells wrote:
> I have been dealing with loss of synchronization on and off for a while
> now. I have captured loopstats and peerstats. The data show what I
> already observed. My clients will synch to my server for a few days
> or so and then reject the server.
>
> The server itself has no problems remaining synched to tock.cs.unlv.edu.
>
> The stability data is interesting. The stability is fairly constant for
> about 5 days then trends up and down wildly and returns to near zero only
> after a reboot. The thing that is odd is that there is any period of
> stability at all. I would guess that unstable clock would be unstable all
> the time.
>
> The other trends I see I think I can explain by system restarts and other
> interuptions. (I am still learning how understand NTP correctly.)
>
> I have imported my trend data into Excel to create graphs. It seems to me
> that I might overlay data, perhaps stability and error, to determine cause
> and effect of my clients rejection of my server. What are the first
> trends that the savvy NTP admins look at to determine the cause of
> problems?
>
> Is there anything I can do to fix a stability problem? Stated another
> way, is stability fixable by changing the system configuration?
>
It would be helpful to see some of the data, knowing the O/S and version
and knowing the version of ntpd that you are running. Also, what's in
the drift file?
Danny
> Thanks,
> Jason C. Wells
>
> _______________________________________________
> questions mailing list
> questions at lists.ntp.isc.org
> https://lists.ntp.isc.org/mailman/listinfo/questions
>
More information about the questions
mailing list