Re: [Zinq-general] Caching SERVFAIL?

Words by Jos Backus [Wed, Mar 25, 2009 at 04:54:40PM -0700]:
> On Wed, Mar 25, 2009 at 06:30:32PM -0500, Mark Johnson wrote:
> [snip]
> > Hi Jos!
> 
> Hi Mark :)
> 
> > http://marc.info/?l=djbdns&m=108560819606350&w=2
> > 
> > Questions, Queries, Posers:
> > 
> > Why the change?  What's with the DNS_T_AXFR?  I'll take a stab at the
> > former and guess maybe that the downside from a cached SERVFAIL
> > (bounced mail, etc) was determined to be worse than the performance
> > benefit of not repeatedly trying to contact horked servers?
> 
> That sounds like a good guess (but I'm no DNS expert). "It seemed like a good
> idea at the time." Per IP is obviously harder to implement.
> 

Acording to rfc2308:

"it MUST be cached against the specific query tuple <query name, type,
class, server IP address>.

I did a naive patch to dnscache that cached just the ip and it greatly
reduced the queries wait. That is until things went wrong (if I recall
correctly serverfailed() only groks NOERROR and NXDOMAIN and on
REFUSED replieѕ for AAAA records I cached the IP as SERVFAIL :|).

-- 
Jose Celestino | http://japc.uncovering.org/files/japc-pgpkey.asc
----------------------------------------------------------------
"One man’s theology is another man’s belly laugh." -- Robert A. Heinlein