On Wed, Mar 23, 2005 at 09:21:38AM +0100, Adrian von Bidder wrote:
> Every hour,
> (1) each server's score is multiplied by 0.95
> (2) if the server is unreachable from some chosen monitoring server,
> score -= 5
> (3) if the offset is >1.5s,
> score -= 2
> (4) if the offset is >0.5s,
> score += -2 * $offset + 2
> (4) if the offset is <=0.5s
> score += 1
> (5) regenerate zone files from all servers with score > 5
> $offset is the offset arrived at by sending one UDP query - I decided that
> just one packet is enough, because the occasional lost packet doesn't
> matter much, it takes 4 lost packets to drop the score below 5 on a good
> server.
2 days ago I go out pool due a ntpd problem (my old problem, resolved
with 4.2.0-8 debian sid, now no more segfault), while I was under 5
point I have an avarage of 660 client connected.
Your system work fine, but somebody around the world have a DNS
configured not very well, my ip appears in pool too. (Don't ask me why,
I've tried about 100 ping and a pair of time my ip is returned).
There isn't a way to say other DNS do NOT cache more than 1 hour?
Regards
Sythos
--
Sythos - http://www.sythos.net
() ASCII Ribbon Campaign - against html/rtf/vCard in mail
/\ - against M$ attachments
_______________________________________________
timekeepers mailing list
[email protected]
https://fortytwo.ch/mailman/cgi-bin/listinfo/timekeepers
