[dns-operations] Can a 1s TTL cause weird failures?

[Mark Andrews]

> Mark Andrews wrote:
> > 	BIND 8 doesn't have query restart.  Upgrade.
> 
> I'm sorry, I don't know what "query restart" is.

	BIND 8 uses UDP query retransmissions to restart the query
	process after fetching glue.  BIND 9 just restarts the
	original query processing.  In the time between retransmissions
	the glue expires.

	You can use a forward zone to work around this but not this
	solution does not scale.  It also requires regular checking
	to follow delegation / nameserver address changes.

> In addition, upgrading 
> to BIND 9 is on our road map but this problem is happening now.  As a 
> large organization we can't just upgrade servers at the drop of a hat, 
> much as I might like to.  :)

	BIND 8 was declared dead 6-7 years ago.  Yes, I know there were
	releases after that, in particular IPv6 support w/ 8.4.x.  None
	of these fixed the fundemental problems w/ BIND 8.
 
> Are you contending that "query restart" is some kind of feature missing 
> from BIND 8 that is causing this problem?

	It was a known problem in BIND 8 that was addressed with BIND 9.

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