[ntp:questions] Re: [Bug 177] Clock stepping messes up frequency. (fwd)

David L. Mills mills at udel.edu
Wed Nov 12 06:21:22 UTC 2003


Ulrich,

I only hate it that I haven't been to beautiful southern Germany in too
long.

My experience here with maxpoll ~17 has been gemutlich. The poll
interval for experimental paths to the local net, to noisy
tick.usno.navy.mil and to time.nist.gov worked just as expected. The
residual jitter at the longer poll intervals increases of course, but
well within the step interval. On some paths the inteval wanders between
12-14 and others 14-17, depending on local temperature fluctuations and
network jitter. Be well advised, however, this was after the
modifications I mentioned in my previous message to this group. We
strive for incremental improvements.

Dave

Ulrich Windl wrote:
> 
> roy at suespammers.org (Roy) writes:
> 
> ...
> > I had set the MAXPOLL up to 17 in an attempt to reduce the load on the
> > NTP servers.  But with that configuration, ntpd sends a whole lot of
> > packets at MINPOLL every time it exceeds the step threshold.  MAXPOLL
> > of 15 seems to be more stable.
> ...
> 
> Stupid as I am, I'd say that ntpd should only increase the polling
> interval if it observes the clock being stable enough. A similar thing
> is true for the PPS. IMHO the hardwired tolerance is far too big,
> making the algorithms increase the polling interval. Dave will hate me
> for saying that...
> 
> Ulrich



More information about the questions mailing list