bind and Beowolf Clusters

[Martin McCormick]

	Does bind take advantage of the parallel redundancy of a
Beowolf cluster running FreeBSD?

	The interest here is redundancy of the master DNS.  It would
be nice to know that, if water flooded a room containing a node or
two, ddns updates could still work and the public wouldn't notice a
thing.

	Some months ago, one of our slave bind servers had been up for
330 days and I was hoping to make a full year or more.  Then, a
maintenance worker accidentally yanked a power cord out of the strip
and, so much for that year.  Sooner or later, stuff is going to happen
no matter how good the OS, hardware or software.  With the kind of redundancy I
am hoping for, an incident like that would be unnoticed by all but
network people and could be repaired the next working day rather than
send the whole brigade over the top right now!!!

	We have had nothing but stellar quality of service with our
bind9/FreeBSD servers, but parallel redundancy would be even more
robust than the almost 100% up time we already have.  It would be what
I call a good answer to those who would say, "What happens if the
master fails?"

	Thank you

Martin McCormick WB5AGZ  Stillwater, OK 
OSU Information Technology Division Network Operations Group