assertion failure bind 9.2.1

Mark_Andrews at isc.org Mark_Andrews at isc.org
Tue Aug 13 17:27:36 UTC 2002


> On Wed, 14 Aug 2002 Mark.Andrews at isc.org wrote:
> 
> > 
> > > 
> > > Strange... our secondary NS died this morning with the following
> > > message:-
> > > 
> > > named[7322]: adb.c:727: INSIST((((name) != 0) &&
> > > (((const isc__magic_t *)(name))->magic == ( (('a') << 24 | ( 'd') << 16 |
>  (
> > > 'b') << 8 | ( 'N')))))) failed
> > > named[7322]: exiting (due to assertion failure)
> > > 
> > > This comes from the line in "adb.c" that says:
> > > 
> > > INSIST(DNS_ADBNAME_VALID(name));
> > > 
> > > Anyone else seen this?
> > 
> > 	It's a know race problem with the threaded build that we are yet
> > 	to track down.
> > 
> > 	Mark
> > 
> 
> Hmm... We're running on OpenBSD 2.9, the bit of configure said (rather
> confusingly)...
> 
> $ ./configure --with-openssl
> checking whether to build with thread support... no
> ...
> checking whether to build with threads... no
> ...
> creating lib/isc/nothreads/Makefile
> creating lib/isc/nothreads/include/Makefile
> creating lib/isc/nothreads/include/isc/Makefile
> ...
> configuring in lib/bind
> ...
> checking whether to look for thread support... yes
> checking for pthread_create in -lpthread... no
> checking for __pthread_create in -lpthread... no
> checking for __pthread_create_system in -lpthread... no
> checking for pthread_create in -lc_r... yes
> ...
> checking for pthread_attr_getstacksize... yes  
> ...
> checking whether to build with threads... yes
> 
> So maybe some bit is built with threads and some isn't??
> 
> Dave
> 
> -- 
> Dave Hill                   <dhill at zeus.com>             Senior SysAdmin
> Tel: +44 1223 525000                                Fax: +44 1223 525100
> Zeus Technology Ltd                                  http://www.zeus.com
> Zeus House, Cowley Road, Cambridge, CB4 OZT, UK
> 

	I should make the two configure's consistant w.r.t. default
	thread support.  However libbind is not used by named.  It's
	entirely seperate.

	Do you have a core dump (i.e. we you running w/o -u).
	If so can you send to bind9-bugs at isc.org a stack backtrace
	and *name.
	
	Mark
--
Mark Andrews, Internet Software Consortium
1 Seymour St., Dundas Valley, NSW 2117, Australia
PHONE: +61 2 9871 4742                 INTERNET: Mark.Andrews at isc.org


More information about the bind-users mailing list