On Jan 9, 2006, at 3:16 PM, Adrian von Bidder wrote:
All five would have to be down before a significant problem would
happen, and even then other DNS servers would still have cached
info...
Of course, if the primary is down, the zone file won't be regenerated
anymore and some of the ntp servers in the rotation right at that
moment
might get traffic problems.
Well, if the primary DNS server or the mechanism which generates the
zone files for it goes down, then we're all stuck using the last-
available information.
People who have set up stratum-1 and -2 timeservers ought to be fine
since they should not be using the NTP pool addresses themselves,
which means that the upper levels of the NTP infrastructure ought to
keep going. And clients which are already in a stable NTP
relationship ought to continue working fine.
New connections to the NTP pools might not be load-balanced properly
if the outage persists for a long time, but it would probably need to
be down for a period of several days before it would make much
difference.
I guess the only workable solution here would be to distribute the
whole
database and generate the zones locally on each nameserver. But
having 5
different authoritative zones at any given time seems ... wrong. I
think
if we're starting to worry about this, we'd better face the fact
that DNS
is not the right tool for what we're trying to do, and start
hacking ntpd
*now*.
What problem are we trying to solve?
--
-Chuck
_______________________________________________
timekeepers mailing list
[email protected]
https://fortytwo.ch/mailman/cgi-bin/listinfo/timekeepers
