[ntp:questions] Re: Dopey selection of server
David L. Mills
mills at udel.edu
Fri Jan 30 17:51:35 UTC 2004
I commented on a question like yours a week ago. Currently I do this at
least once per month. There is a lot more going on than you suspect,
including recent history, source jitter over time, hysteresis, etc. Look
at the peer state variables and note the dispersion, synchronization
distance and jitter. All is revealed at the NTP project page in the
architecture briefing. So far, very few folks have had the patience and
interest to actually do that.
I would be absolutely delighted if you read that and found the actual
implementation violated those principles. Meanwhile, start ntpd with two
-d command line switches and watch carefully what the algorithms do.
Again, I would warmly welcome a critique of those algorithms, but only
if the watcher knew what they were trying to do. Should someone point
out something wrong in either the principles or implemenation, I would
change the code in a heartbeat.
Christopher Burian wrote:
> Can anyone explain this:
> ntpq> pe
> remote refid st t when poll reach delay offset
> +cs.columbia.edu 184.108.40.206 2 u 725 1024 377 16.819 -7.688
> +filbert.cc.colu 220.127.116.11 2 u 704 1024 377 22.523 -4.356
> *cudns.cit.corne 18.104.22.168 2 u 656 1024 377 60.335 13.325
> Instead of rejecting the one that is off the most from the others, it is
> selecting that wayward server, despite it having the worst jitter as well.
> Is the selection based on a longer time frame, and I just happened to pick
> a bad time to look at how it's doing?
More information about the questions