On 1/25/07, [EMAIL PROTECTED] <[EMAIL PROTECTED]> wrote:
> Of course, you could implement a more pool specific solution ( just doing
> a new DNS lookup at the pool when one server goes down ) but I would be in
> favor of a more generic solution ( active and spare ntpd server in the
> config )
>
I would suggest behavior that increases maxpoll for an *unreachable*
server to at least 2**16 seconds (about once a day). It could then
re-issue DNS lookups for unreachable servers after say 2**18 seconds
of a server being unreachable.
The above changes would drastically cut down on traffic to dead
servers. The maxpoll for an unreachable server could be left at 2**16,
which would still leave us with 1/16 the traffic we see now. Also, the
behavior change is so benign that it would likely not be encountered
except by badly configured systems or when a server really did go
offline for an extended period. The DNS re-lookup behavior would help
the pool, and have no material effect for non-pool systems. No
confiugration file changes would be required.
> actives and spares is a well known concept although
Having a "spare" keyword, similar to "prefer", wouldn't break any
existing configurations file. So this too seems like a decent idea.
For every active server that's unreachable for two maxpoll periods,
one "spare" is selected. The Microsoft Windows Time Service actually
has this behavior as a registry option.
--
RPM
=========================
All problems can be solved by diplomacy, but violence and treachery
are equally effective, and more fun.
-Anonymous
_______________________________________________
timekeepers mailing list
[email protected]
https://fortytwo.ch/mailman/cgi-bin/listinfo/timekeepers
