Parent zone rejected on problems with slave zone

[Barry Margolin]

In article <bv6hok$3t9$1 at sf1.isc.org>, gorf_55555 at hotmail.com (Peter L) 
wrote:
> > > that it slaves for.  They are not delegated, but we pull the zones as
> > > slave.  Ocassionally, when one of these zones' NS goes off line or
> > > whatever, this child zone will expire.  Upon that event our root
> > 
> > Maybe you should increase the expire times in these zones, so they don't 
> > expire so quickly.  I recommend at least a week.
> > 
> 
> 
> My point is that an expired slave should not affect the parent.

Since the expired zone is a subdomain of the parent, its NS records are 
implicitly included in the parent.  This may be what's causing the 
problem.  However, I don't think I've ever seen this behavior on any of 
the BIND 8 servers I've managed.

> > > nameserver for corpx.com says " master zone "corpx.com" (IN) rejected
> > > due to errors".  As soon as we either correct the connection to the
> > > slave zone and get a good zone transfer, or comment out that slave
> > > zone in named.conf, the root server is fine again.  What is going on
> > > here?
> > 
> > It would help if you told us what the errors are.  They should be the 
> > log messages preceding the one that says "rejected due to errors".
> 
> 
> I thought I was clear about expiration... pasted below.

You were clear about the expiration, but I expected there to be another 
error message saying what's wrong with the parent zone that caused it to 
be rejected.

> Jan 13 13:23:34 jan   root at clpr1:/var/tmp/BIND/src/bin/named
> Jan 13 13:23:34 jan named[2860]: [ID 295310 daemon.notice] Ready to
> answer queries.
> Jan 13 13:23:36 jan named[2860]: [ID 295310 daemon.notice] slave zone
> "cadc.corpx.com" expired
> Jan 13 13:23:36 jan named[2860]: [ID 295310 daemon.error] master zone
> "corpx.com" (IN) rejected due to errors (serial 200312230)
>

-- 
Barry Margolin, barmar at alum.mit.edu
Arlington, MA