Bind 9 becomes unresponsive

[Mark Andrews]

> I just upgraded our bind 8 configuration to bind 9.3.1.  Bind will stop
> answering quieries with the following error in syslog:
> 
> named[27140]: [ID 866145 daemon.crit] mem.c:1607:
> INSIST(mpctx->allocated > 0) failed
> 
> This is running on a Solaris 9 server that is way beefier than what
> should be needed for DNS.  It's a Sunfire V240, 8G of ram, dual
> processor machine.  I have bind running in a chrooted environment, and
> I built it with gcc version 3.4.1.  I have run named-checkconf against
> my named.conf, which resulted in no errors.  I have also run
> named-checkzone against my zones and again it resulted in no errors.
> The server will stop responding at random times, and initially seems to
> perform normally.
> 
> Any ideas?

	There was a attempt to return more objects to the pool than
	had been allocated.

	Errors like this should be reported to bind9-bugs at isc.org
	along with a stack backtrace, compiler used (including version)
	options used to compile named, location where the executable
	and core dump can be fetched from and anything else you feel
	is useful.

	Mark
--
Mark Andrews, ISC
1 Seymour St., Dundas Valley, NSW 2117, Australia
PHONE: +61 2 9871 4742                 INTERNET: Mark_Andrews at isc.org