On 2/21/2010 8:25 PM, Doug Barton wrote:
On 02/20/10 08:54, kalpesh varyani wrote:
Thanks Dave for pointing this out.
the first server did not fail, it behaved as per its configuration.
But for a stub resolver, which cannot follow referrals, isnt it logical
for it to detect referrals and move on to the next name server in the list?
No. What you're describing is the function of a resolving name server,
not a stub resolver. If you think that the stub resolver should behave
differently you need to take that up with your OS vendor.
Stub resolvers are described briefly in RFC 1034, Section 5.3.1. The
verbiage in that paragraph "The user also needs to verify that the
listed servers will perform the
recursive service" clearly assumes that all of the resolvers used by the
stub support recursion. The behavior of a stub resolver when receiving a
referral (non-recursive) response is *undefined* by the standards, and
something to be avoided.
As Doug pointed out, if the resolver is sophisticated enough to deal
with non-recursive queries, then that goes beyond what is expected of a
"stub" resolver, described in the RFC as typically "a PC which lacks the
resources to perform the resolver
function".
- Kevin
_______________________________________________
bind-users mailing list
bind-users@lists.isc.org
https://lists.isc.org/mailman/listinfo/bind-users
