This site has fiddled with their DNS server to perform this trick for
load balancing. Presumably, all hosts that are identifed by these
multiple addresses have exactly the same content. And since you
back off 60 seconds to prevent excess load on someone's server,
no problem exists on your part that I can see.
I think it is strange that the DNS information is not getting cached along
the way. Usually, this multiple address trick works on a macro
rather that micro scale, meaning that an address is cached somewhere,
and the later gethostbyname requests are served by the (non-auth|caching)
server. If you are making a DNS call to the auth server for every
gethostbyname(), there exists an inefficiency on your part.
Bryan Hackney
bbh@xenodata.com
----- Begin Included Message -----
From: Aaron Nabil <nabil@teleport.com>
...
I was testing the lock manager today and was getting (what I
thought were) erratic results until I realized that the site I
was testing against was returning multiple A records (IP addresses).
...
----- End Included Message -----