David,
Hint: Look at the flash bits and clock discipline time constant (rv
hpoll) for the associations involved. I bet you find 10 on both machines
and then something else has clamped down on the 64-s poll.
Dave
David J Taylor wrote:
David L. Mills wrote:
David.
Depending on phasing and poll interval, it might (and does) happen
that during a poll interval one peer will see none, one or two polls
from the other peer. That is normal and does not affect the
synchronization function. Sometimes as the poll interval ramps up
more than one interval may go by without an arriving packet. In the
expected case where the peers are low in the stratum forest and the
highest reliability is needed, it probably is a good idea to clamp
the maxpoll to the default minpoll.
Dave
Dave,
Thanks for that. I can see that the phasing could make the numbers 0, 1
or 2 polls, just depending. What I'm not quite so clear about with peered
servers is how a server can show a poll interval of 1024s against an
external server, and yet 64s against a peered server which is just
starting up, and therefore has an interval of 64s against its own external
servers.
It probably really means I don't understand exactly peering works! "One
server can set the time of the other" is what I think I know.
73,
David
_______________________________________________
questions mailing list
[email protected]
https://lists.ntp.isc.org/mailman/listinfo/questions
