Re: [hylafax-users] Enterprise Hylafax

[Darren Nickerson <darren@dazza.org>]

>>>>> On Wed, 6 Sep 2000, "Jay" == Jay R. Ashworth wrote:

  +> Please note that this solution will accomplish load-balancing for you, 
but it
  +> will not give you the automatic failover I believe you were looking for in
  +> your initial query. Round-robin is fundamental to DNS . . . when there is 
more
  +> than one Resource Record (RR) for something then the DNS rotates the 
answers
  +> it gives. You'e get (roughly) a 50/50 split in answers. Beware of 
nameservers
  +> which will cache that answer for awhile though!

  Jay> Well, in the environment of most HF systems, the nameserver being
  Jay> asked will be the *authoritative* nameserver.

In a lot of cases that will be true. So cach may not be such a problem. 

  +> Anyway, If you're looking for enterprise-class clustering/failover, 
HylaFAX
  +> doesn't presently know how to do this. Sorry!

  Jay> Will the client(s) not try more than once if they get more than one A
  Jay> record back from the nslookup?  If not, this seems like an easy
  Jay> feature to add...

The client does not run 'nslookup', rather it doed a DNS lookup, which returns
a single A RR, at random. You don't get two answers - that's the whole point.

-d




____________________ HylaFAX(tm) Users Mailing List _______________________
 To unsub: mail -s unsubscribe hylafax-users-request@hylafax.org < /dev/null