Which nameservers am I using?

[David Bandel]

On 5/29/05, Michael Hipp <Michael at hipp.com> wrote:
> David Bandel wrote:
> > On 5/29/05, Michael Hipp <Michael at hipp.com> wrote:
> >
> >>Net Llama! wrote:
> >>
> >>>207.191.50.10 is borked.  If i try to use it, i'm not gettting anything
> >>>back either.  Where did you get that one from?
> >>
> >>It's the one for the network this server will be attached to when it is
> >>delivered to the client. I presume it's only accessible internally.
> >>
> >>But that's not the point. Why does it just stop after the first one? Why not
> >>try all three listed servers?
> >>
> >
> >
> > Because the server didn't fail, _did_ reply (just not a usable reply).
> >  It didn't give a recursive lookup.  The IP you came from is probably
> > part of a view that has "recursive no;" as an option.
> 
> Ok. What I'm understanding is that remote machine that didn't want to answer
> me in effect said "I'm not going to answer and you're not allowed to ask
> anyone else either." Why does the algorithm allow such a thing? Why wouldn't
> my nslookup/dig just keep going down the list until it got a usable answer
> rather than slavishly obeying such an order?

The resolver libraries are written in such a way that you either get
an answer (in which case they stop trying) or you don't (in which case
they continue on to the next nameserver).  The libraries can't (and
shouldn't) be written to second-guess an answer they receive (but
you're welcome to take a crack at rewriting them to do so -- I just
won't use them).

Your question should probably be:  Why did the nameserver respond at
all if it wasn't going to do the job?  Well, perhaps because if you're
not on their network, they don't want you trying to use their server. 
It is theirs after all.  Their server, their rules.

Ciao,

David A. Bandel
-- 
Focus on the dream, not the competition.
            - Nemesis Air Racing Team motto