Is dig really recursive?

Kevin Darcy kcd at daimlerchrysler.com
Thu Feb 8 04:40:10 UTC 2001


"dig" itself isn't recursive, but by default it sends recursive queries and
expects the nameserver it uses to recurse to get the answer to the query.

When you say "no information" do you mean dig displayed a valid header and
0 answers, an NXDOMAIN response, or did dig just time out? I could sort of
understand a timeout, since BIND 8 -- assuming that your local nameserver
is BIND 8 -- doesn't have query restart; this basically means that in
certain circumstances it'll just drop a query that it has partially
resolved and expect the client to retry, upon which it completes the name
resolution.

You could always verify whether it was just a fluke or not by mimicking the
actions of a recursive resolver (use the +norec flag with dig and follow
the delegations down from the root).

Note that the default retry parameters of dig don't necessarily exactly
mirror those of your system resolver, so maybe your system resolver might
be able to resolve the name even if dig cannot.


- Kevin
atom at suspicious.org wrote:

> i'm not sure if this is a fluke, but i was digging for a 4th level .us
> domain. (someplace.sometown.state.us)
>
> i was getting back the dig results with no information. after digging
> for other .us domains, i couldn't find one that didn't work. (all other
> .us domains worked fine)
>
> then, after digging backwards ("state.us", then "sometown.state.us",
> then the full "someplace.sometown.state.us") it worked fine, but i'm not
> sure if the answer was cached, or if something might have just been down
> when i was doing it the first time????
>
> so... if i dig for "machine.town.state.us", should it "just work", or
> should i expect to have to fight with it?
>
> ...atom
>
> Sent via Deja.com
> http://www.deja.com/





More information about the bind-users mailing list