a question about BINDv8.2.3

Kevin Darcy kcd at daimlerchrysler.com
Thu Apr 18 22:58:18 UTC 2002


angel wrote:

> hi,everyone
> I have a sun workstation,it is a E450.I compiled BINDv8.2.3 on it in
> May 10,2001.Recently ,I found some strange things about it .If I had
> started the process of named about 20 days,some zone names would not
> be resolved .I used "nslookup" and "dig",the result was TIMEOUT.At
> last ,I had to kill the process and restart the process,then the
> problem would be disappear.This phenomenon appeared serveral times.I
> wonder wether it is a bug or a hacking track.If someone know about it
> ,please help me to resolve it .

Well, if it only affected "some zones", then I'd be more likely to chalk
this up to bad delegation information for those particular zones.
Restarting the named process has the incidental effect of clearing the
cache and forcing the nameserver to get delegation information from the
nameservers of the parent zone (e.g. the TLD servers), which is
oftentimes more usable than what is cached from the delegated
nameservers. To tell for sure, I'd need to actually look at the zone
data for the zones in question.

If the nameserver stopped answering *all* queries, then I'd be more
inclined to think that this was a resource issue or a bug in named
(probably one that's been fixed in later releases -- hint, hint).


- Kevin




More information about the bind-users mailing list