[ILUG] dns problem

[Peter McEvoy]

On Wed, Apr 06, 2005 at 11:09:54AM +0100, Declan Grady wrote:
> I have a debian woody firewall running caching only nameserver (bind9).
> outside world is on eth1, and local lan on eth0
> I also have a redhat7 mailserver on the eth0 lan.
> 
> Most things seem to be working ok, but ...
> I currently have a load of email (well, not a load, only about 20 or so) 
> backed up
> for one destination domain, and mailq tells me ' host name lookup failure'

What MTA is this? I thought mailq only showed what mail was on the
queue waiting to be delivered.


> However, if I go to the firewall, and  'dig' that domain it does resolve ok.
> 
> I dont have dig or nslookup on the mailserver, so cannot  do the same 
> dig/nslookup there.

I would install dnsutils or whatever the equivalent redhat package is.

 
> on the mailserver, my  /etc/resolv.conf reads :
> -----------------------------
> domain
> nameserver 192.168.x.x
> ---------------------------
> where 192.168.x.x is the IP of the firewall
> 
> Routing seems ok, and fetchmail works from the mailserver to get my odmr 
> email ok. (also using dns to resolve the hostname)
> Incoming mail seems ok .. some dns failures for the same domain name I 
> have problems sending to ... possibly an error in the config of that 
> domain's dns 'setup' ?
> 
> How do I go about finding what the problem is ?

You can toggle verbose logging with "rndc querylog" , but I've found
restarting bind9 can solve a lot of problems.
It might also help to mention what the domain is, so people can
ascertain whether there are existing mx/a records.

-- 
Pete