[ntp:questions] How should an NTP server fail?
David L. Mills
mills at udel.edu
Fri Jun 11 03:24:49 UTC 2010
The design is to select only one sample in the filter. A read of the
code and inspection of the debug trace confirms that it does exactly
that and results in the behavior I described. Also, from my previous
message to Miroslav, he may be using an older release version which had
a most dubious pedigree. Only recently has the release version became
closely aligned with the development version.
>On Jun 9, 7:41 pm, "David L. Mills" <mi... at udel.edu> wrote:
>>The number of samples in the clock filter has nothing to do with the
>>selection process, nor whether the peer is the system peer or not. The
>>selection alogorithm doesn't even know how many samples are in the
>>filter, only that the filter candidate that is used has least delay. The
>>selection metric includes that and the dispersion at the measurement
>>time, plus the dispersion increment since then. When two or more servers
>>are configured at substantially the same delay, the client may
>>occasionally hop from one to the other depending on these factors,
>>although there is a anti-hop scheme that discourages this unless there
>>is a substantial difference.
>Isn't that the point I was trying to make. You know the old saying, in
>theory, there is no difference between theory and practice. In
>practice, there is.
>My point is that we have the design, the implementation and the
>reality. You say it works one way by design and then examine the code
>to confirm it. But the user reports something very different. Either
>the user is mistaken in the report, is using different code, or there
>is a bug. In all three of these scenarios I would think that that the
>bug report in bugzilla would have been the appropriate place to
>You say above that the number of samples cannot make a difference. The
>user says that it does. We need to reconcile these two things.
>questions mailing list
>questions at lists.ntp.org
More information about the questions