[ntp:questions] Remaining synced on an unsynchronised peer?

unruh unruh at wormhole.physics.ubc.ca
Thu Dec 3 18:16:04 UTC 2009


On 2009-12-03, michael.buetow at comsoft.de <michael.buetow at comsoft.de> wrote:
> I'll try to explain why I do not think that covers it.
>
> In my case, the prefer peer (ntp1) is unreachable.
> Even if it were never discarded (I assume that's what the immunity
> idol means), my host1 and host2 should not use each other as
> synchronization source.

Apparently ntp has a command option called noselect, which tells ntp
never to select that host as a source. It is useful for using A to keep
track of B by periodically sending B packets to see what B's time is,but
never to use it as a source. It sounds like that is what you want. (I
have heard this as a rumour in a chrony discussion, so will have to let
you look up the details-- sorry)

>
>> Mitigation Rules
>> "As the selection algorithm scans the associations for selectable
>> ? ?candidates, the modem driver and local driver are segregated
>> ? ?for later, but only if not designated a prefer peer. ?If so
>> ? ?designated, a driver is included among the candidate population."
>
> Problem is, my prefer peer (to host1/host2) is neither a modem nor
> local driver.
> It is a genuine external source.
>
>> "If the prefer peer is among the survivors, it becomes the system
>> ? ?peer and its clock offset and jitter are inherited by the
>> ? ?corresponding system variables. Otherwise, the combining
>> ? ?algorithm computes these variables from the survivor population."
>
> The prefer peer indeed remains the synchronisation source, despite
> being unreachable (!).
> This is what I find buggy.
>
>> The minsane Option
>> "The minclock threshold
>
> Thanks, didn't use these options before. Maybe they will become handy
> at some time, but as far as I can see they will not mitigate my
> immediate problem. If I cut the LANs, I expect to have no survivors.
> Shouldn't my preferred peers have been eliminated as unfit due to no
> reachability in ntp_proto.c:peer_unfit() (TEST13) ?




More information about the questions mailing list