Slave's lack of NS records cause zone file dump

SilentRage bind-users at
Fri Oct 22 05:19:27 UTC 2004

I have a recurring problem, which may be considered a bug, but I'm looking for advice on how to handle this.  I have a client who added a slave to the dns server.  The problem is the master server's zone lacks NS records.  The transfer occurs fine and the slave serves records despite the lack of authority records.  However, if the dns server needed to be restarted for some reason (rare occurance luckily) then BIND complains when it loads the slave zone.  It dumps the zone to a filename like db-00016356 in the working directory and returns "Server Failures" to all new requests.  A retransfer of the zone is necessary to cause it to work again.

Problems with this behavior:

1) I don't like the fact that it starts serving from a file other than the zone file specified in the config.  Not to mention I dislike the clutter.
2) I don't like the fact that it starts working perfectly, and later fails.  It should either never work in the first place, or always work.  Whether it accepts authority-less zone transfers should be configurable (something similar to check-names).

What can I do about this besides monitor logs and email those clients who fall under this admittedly unusual problem?


More information about the bind-users mailing list