No notify-refresh: bug or feature.

Mark Andrews Mark_Andrews at isc.org
Tue Mar 18 02:09:56 UTC 2008


> > > then, see above, let's just remove the "include new RR in NOTIFY-Q" logic
> ,
> > > and simulate that removal in our notify responders.
> > 
> > 	That's throwing the baby out with the bath water.
> 
> i think we don't know enough, by a long shot, to define correctness here.
> 
> > 	Just allow the record to be used as a hint about whether
> > 	refresh processing should be performed or not.  Leave it
> > 	up to the implementation to decide what parts of the hint
> > 	it listens to.
> > 
> > 	A implementation MAY use the answer section to decide if
> > 	refresh query processing needs to be initiated.
> 
> that's possibly correct if the NOTIFY-Q TYPE is SOA, but possibly not, and
> certainly not otherwise (since only SOA has associated "refresh query
> processing.")

	NOTIFY for anything other than SOA is undefined.  If/when such
	requests make sense then the behaviour should be defined.  NOTIMP
	is currently the correct response code.

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


More information about the bind-workers mailing list