Strange Zone Errors :(

[Mark.Andrews at nominum.com]

> 
> Okay, let me see if I have this straight:
> 
> 1. nameservers get the "real" NS'es (stargate and yoda in this case)
> 2. the corresponding A records happen to expire from the cache earlier than t
> he
> NS'es themselves
> 3. these A records wouldn't be obtainable from the parent servers because the
> y,
> i.e. the parent servers, would have no way of knowing that they, i.e. the
> A records for the "real" NS'es, were necessary glue records; and the delegati
> on
> NS'es (dns1 and dns2) from the parents would never replace the "real" NS'es a
> s
> long as they are cached because the "real" ones are more credible, right?
> 4. Therefore deadlock occurs (at least until the NS'es expire from the cache 
> and
> the cycle starts all over again).
> 
> Did I get it?
	
	Yes.
> 
> I wonder if there's an easier-to-comprehend way of explaining this, so that i
> t can
> be promulgated and folks will understand that it's more than just a
> "hygenic" issue...
> 

	Not really.  One of these days I'll get around to having the
	child's nameservers complain if the delegation does not match.

	Patches welcome.

	Mark
> 
> - Kevin
--
Mark Andrews, Nominum Inc.
1 Seymour St., Dundas Valley, NSW 2117, Australia
PHONE: +61 2 9871 4742                 INTERNET: Mark.Andrews at nominum.com