named process dying on rndc reconfig

[Mark Andrews]

> I observed the same effect on two different nameservers yesterday.
> Each was BIND 9.3.4 running in a Solaris 10_x86 "zone".
> 
> I had modified a file included from named.conf to remove a particular
> IP address from many "masters" clauses for slave zones. On "rndc reconfig"
> the named process softly and suddenly vanished away. There were no
> messages (at severity "error" or higher). On restarting, BIND showed
> no dislike of the modified configuration files.

sounds like

2115.   [bug]           'rndc reconfig' could trigger a INSIST if the
                        number of masters for a zone was reduced. [RT #16444]

> 
> As I have made similar modifications many times before without any
> such problem, I wonder whether it is something specific like "removing
> an IP address from a masters clause while an attempted zone transfer
> from that address is in progress". The reason the IP address was being
> removed was that the machine addressed by it had become u/s, and the
> slowly timing out zone transfer attempts were slowing down others, so
> certainly there were always such attempts going on.
> 
> Has anyone observed anything similar to this?
> 
> -- 
> Chris Thompson
> Email: cet1 at cam.ac.uk
> 
> 
-- 
Mark Andrews, ISC
1 Seymour St., Dundas Valley, NSW 2117, Australia
PHONE: +61 2 9871 4742                 INTERNET: Mark_Andrews at isc.org