Failover network strategy

Jeff Lightner jlightner at water.com
Thu Sep 21 15:47:42 UTC 2006


What is the best way to deal with failover networks served by DNS?
Say we have a range of IP addresses assigned to us from the outside
world and existing DNS servers have been setup to resolve the names to
the appropriate places.  (Registrar setup done, reverse lookups etc...
all done and working, NATting done to the real internal IPs of the
various destinations.)

We also have another range of IP addresses assigned to us with separate
switches that would allow NATting these addrsses to the same real
internal IPs of the various destinations.

We could of course just set up different DNS servers on this other range
that only know this other range of IPs.  The downside being we'd have to
either have them registered all the time and turned off (allowing the
possibility of needless timeouts on queries) or not registered until we
need them which means there could be a delay between the time the
primary range failed and the second set of DNS servers were seen by the
outside world.

I don't think we're the first to look at this so was wondering how other
people do it.

Jeffrey C. Lightner
Unix Systems Administrator
DS Waters of America, LP
678-486-3516





More information about the bind-users mailing list