Bind 9.7 - sanity check or a bug

[Din Jo]

Please help me to understand the following behaviour of Bind 9.7.

Consider this:
                       NS      server1.test.com.
                        NS      server2.test.com.
server1                 A       10.0.0.1
server2                 A       10.0.0.2

case 1:
# nsupdate
> server 127.0.0.1
> update delete server2.test.com A
> update add server2.test.com 2222 A 10.0.0.2
> send
> quit
No issues in above case.
case 2:
# nsupdate
> server 127.0.0.1
> update delete server2.test.com A
> send
> update add server2.test.com 2222 A 10.0.0.2
> send
update failed: REFUSED
> quit

syslog will show: "updating zone 'test.com/IN': update rejected: post update
name server sanity check failed"

In case 1 it is not giving any error. Why it is giving "sanity check error"
in case 2 only?
If this is because "NS records without A records are not allowed", then it
should not allow us to delete the NS A record also. While deleting NS A
record also it should have given the sanity check error.
Is it a bug in Bind 9.7? This issue is not seen in Bind 9.5.
-Dinmadh
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://lists.isc.org/pipermail/bind-users/attachments/20110128/afca490a/attachment.html>