CNAME and other data?

Simon Waters Simon at
Tue Oct 29 11:56:45 UTC 2002

Kevin Darcy wrote:
> Hmmm... Looks like it is coming from the cache. Moreover,
> seems to be some sort of load-balanced cluster -- I'm seeing the TTL values
> alternate between greatly different values on successive queries. Maybe one of
> the nodes in the cluster isn't actually BIND 8.3.3, but may have been fudged to
> say so in order for the answer to the version query to be consistent (???)

A caching BIND 9 will give similar answers for the same domain.

The short answers is GIGO, the domains DNS is a complete mess,
and that it produces any answer is more luck than judgement.

Although I guess the algorithmn for answering DNS queries says
throw away the glue NS records from SLIST, it doesn't say drop
the A records from cache and don't use them in an answer if you
are a recursive server which is what I think is happening.

It gets very confusing when looking at such broken domains.

More information about the bind-users mailing list