DNS Problem: Help!

Mike Diggins diggins at McMaster.CA
Sat Nov 16 23:16:49 UTC 2002



On Sat, 16 Nov 2002, Simon Waters wrote:

> Mike Diggins wrote:
> >
> > The IP address for their
> > name server NS.CGOCABLE.NET comes back as 24.226.1.11 which is wrong, it's
> > not even a name server. It should be 24.226.1.94! If I restart named it
> > sometimes picks up the correct address. I just restarted both my name
> > servers and find that they have, once again, an incorrect address.
>
> You must run BIND 8 or earlier, as the correct thing to do would
> be to ignore the whole domain like BIND 9 should ;-). Upgrade!
>
> The wrong IP address is because of stale glue. The gtld-servers
> think it is 24.226.1.11, and pass this stale data along.
>
> > Is there anyway to force our name servers to lookup
> > this information again without restarting named? Any help would be
> > appreciated.
>
> Of the three IP address for the name servers for this /24 two
> are lame and one is not a name server.

Could you tell me how you discovered this?

> If the zone is further delegated from the ISP to the customer
> servers, you could slave the delegated zone, or mock up a
> working zone for the IP addresses if it is just to satify the
> requirement to reverse lookup correctly, and thus bypass the ISP
> config, otherwise I'd say it is the customers problem if there
> ISP is can't run their DNS correctly.

I may have to do that if the ISP can't get this fixed. It's a funny
problem, normally I wouldn't expect it to be a major problem but it seems
to cause problems with Solaris. Have a look at the output of a ping (with
-s) from a solaris machine to a live host when I point resolv.conf to a
non-existant DNS server. Not only can't it lookup the address it refuses
to accept the echo reply for 300 seconds. I've sniffed the line and the
echo replies are there, Solaris just doesn't "see them".

diggins at newblack<~># ping -s 130.113.128.1
PING 130.113.128.1: 56 data bytes

{pause of about 300 seconds}

64 bytes from 130.113.128.1: icmp_seq=0. time=0. ms
64 bytes from 130.113.128.1: icmp_seq=1. time=299052. ms
64 bytes from 130.113.128.1: icmp_seq=2. time=298052. ms
64 bytes from 130.113.128.1: icmp_seq=3. time=297053. ms
64 bytes from 130.113.128.1: icmp_seq=4. time=296053. ms
64 bytes from 130.113.128.1: icmp_seq=5. time=295053. ms
64 bytes from 130.113.128.1: icmp_seq=6. time=294053. ms
64 bytes from 130.113.128.1: icmp_seq=7. time=293053. ms
64 bytes from 130.113.128.1: icmp_seq=8. time=292053. ms
64 bytes from 130.113.128.1: icmp_seq=9. time=291054. ms
64 bytes from 130.113.128.1: icmp_seq=10. time=290054. ms
64 bytes from 130.113.128.1: icmp_seq=11. time=289054. ms

etc (down to normal ping response time)

-Mike




More information about the bind-users mailing list