wierd bind problem.... queries not answered...

Jeffrey C. Albro jeff at velvet.antistatic.com
Fri Dec 22 17:31:33 UTC 2000



I'm having a really wierd problem...  When my Debian 2.2 box running bind
8.2.2 p7 is rebooted, named restarts and sends out priming queries, and
never gets any answer.  So I would type something like:

nslookup debian.org

and I would get:

Server:  localhost
Address:  127.0.0.1

*** localhost can't find debian.org: No response from server

Now, I'm monitoring all this with

tcpdump -nx -s 128 ip and port 53

And I can see the requests go out, but nothing ever comes back in, from
either the forward servers or the fail-over root servers.  I can ping
these servers successfully while this is going on.

Now THIS is where it gets strange.  If I restart bind with

/etc/init.d/bind stop
/etc/init.d/bind start

(and no other changes)

SOMETIMES it will start working again and everything looks as it should,
including the tcpdump output.  On the second restart it usually works.  
We only have a packet firewall, not a stateful one, so that's not the
issue...

(we are making the request on a non-priveldged port)

The only thing I can think of is sending mal-formed queries out, but even
those should get some kind of response.  The other possibility is that
this is a hardware problem, as this first appeared after a power outage
during a lightning storm, but two identical machines (va linux 2 by
2s) are doing the exact same thing.  I also tried switching from my
home-compiled verions to the offical Debian binaries.

Any ideas?  Need more information?

Thanks!

-Jeff





More information about the bind-users mailing list