[ntp:questions] Odd (mis)behavior when reference clock fails

Unruh unruh-spam at physics.ubc.ca
Wed Sep 24 05:35:38 UTC 2008

Steve Kostecke <kostecke at ntp.org> writes:

>On 2008-09-23, Unruh <unruh-spam at physics.ubc.ca> wrote:

>> That is precisely what ntp is supposed to do for you. 
>> Unless you are suggesting that the clock drift is greater than 500PPM.

>Unidirectional drift and periodic clock steps are a good indicator that
>the clock drift exceeds ntpd's correction capabilities. In these cases
>adjusting the tick to bring the clock drift within +/- 500PPM is
>sometimes the solution.

>As suggested elsewhere in this thread stopping ntpd, deleting the
>drift.file, and starting ntpd is another potential solution.

Overdrifting does not seem to be the problem as in his recent post. His
drift of 5ms in a day is only .1PPM. ntp should be able to handle that
easily and it indicates some other problem. And that is less than one lost
tick per day, if that were theproblem. Strange.

More information about the questions mailing list