Different DNS resolutions from 2 servers

Mark_Andrews at isc.org Mark_Andrews at isc.org
Thu Oct 24 04:25:36 UTC 2002


> 
> Thanks for the response.  Its interesting, but I don't really buy it.
> 
> First, I don't control the old name servers.  I cannot do anything about
> them.  I'm only running a recursive server for this problem.
> 
> Secondly, there has to be a mechanism whereby the recursive server checks
> against the root servers when the TTL for the authoritative NS expires.
> Otherwise, ISPs could hold domains hostage when a customer moves.  Also,
> all of my BIND 4.93 have handled this correctly, its just the BIND 9.2.1
> that are having problems.  And its not all of the 9.2.1 servers, only
> about 10% (I manage over 60 of these) of them are reporting the old NS
> information.

	BIND 4.9.3 (4.93 doesn't exist) didn't replace the record
	if it already existed in the cache so the TTL didn't get
	refreshed.

	If a ISP wants to hold a zone captive you sue them however
	most ISP's really don't want the extra traffic that results
	from this which is also unfunded.

	Well those ones that have the old data were talking to the old
	servers.  Flush their caches.
	
	Mark

> 
> Any other ideas?
> 
> Donald E. Foss
> ###############################
> 
> On Thu, 24 Oct 2002 00:12:33 +0000, Mark_Andrew wrote:
> 
> 
> > 	Turn off the old nameservers or make them slaves and then turn them off
> > 	once all the TTL referencing them have expired. At the moment the old
> > 	nameservers are refreshing the NS RRset everytime they are queried.
> > 
> > 	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
> 
--
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