[ntp:questions] strange behaviour of ntp peerstats entries.

David Woolley david at ex.djwhome.demon.co.uk.invalid
Sun Feb 3 17:34:39 UTC 2008


Unruh wrote:

> night and getting myself confused again. I do not think my comments about
> the aging are correct.

It looks like ones older than the Allan Intercept are aged, but they are 
defined in such a way that they are always worse choices than ones 
before the Allan Intercept.
> 
> I think that the process is --look though the last 8 samples for the one
> with the smallest delay, If that is the the latest one, use it and go to

If you mean this test:

         if (peer->filter_epoch[k] <= peer->epoch) {

peer-> epoch is only updated if it succeeds, so the test isn't against 
the latest, but against the latest one that was actually used.




More information about the questions mailing list