[Pdns-dev] Possible ANY, CNAME & cache bug in pdns_recursor ?

[Leen Besselink]

On Fri, Sep 22, 2006 at 11:55:14AM +0200, bert hubert wrote:
> On Fri, Sep 22, 2006 at 11:53:54AM +0200, Leen Besselink wrote:
> > Possible I've found a bug in pdns_recursor, it will return a different
> > answer after I've requested more information.
> 
> That is correct, and as things should be. 
> 

OK, good. :-)

> Please see http://psg.com/lists/namedroppers/namedroppers.2006/msg00456.html
> for details.
> 

Hmmm, not what I would have expected though.

> Thanks for the report! I'm pondering your other ticket on hide-ttl.
> 

I was pondering of adding a note to the bug-report.

Although it might be usefull to some, it's looks like we won't be needing a change
to powerdns-recursor afterall.

As our NAT-firewall (which is also running a recursive DNS) is going to be
moved to OpenBSD. And as powerdns-recursor can not run on OpenBSD at this
time, we'll be using DJB's dnscache for the moment.

Although it's upstream recursive-DNS-servers will probably be powerdns-recursors.

I do however hope OpenBSD will add swap_context and friends soon. :-)

We use hidettl because you can revisit a website at it's new location when you are
doing authoritive DNS changes on a website you visited just a few moments before that.

We have a 'webinterface' which you can visit which just kills dnscache (and a new
one is spawned immediately) to flush the cache of the internal network.

Negative answers are cached however. This is a feature/precaution, not a bug.

> 	Bert
> 
> -- 
> http://www.PowerDNS.com      Open source, database driven DNS Software 
> http://netherlabs.nl              Open and Closed source services
>