[Pdns-dev] Is this a bug ?

Leen Besselink leen at wirehub.nl
Wed Feb 11 08:45:04 CET 2009


bert hubert wrote:
> On Wed, Feb 11, 2009 at 07:58:54AM +0100, Leen Besselink wrote:
>> An old version of PowerDNS-recursor and a djbdns all seam to return something like this (the same after a restart (I checked the djbdns)):
> 
> This is a bug in BIND 9.2.2.
> 

yup, even fpdns seems to confirm that (and it usually is kind of behind the
times when it's finally packaged, etc.):

fingerprint (209.62.177.21, 209.62.177.21): ISC BIND 9.2.0rc7 -- 9.2.2-P3

> If you have a very recent 'dig' from BIND, you can do this:
> 
> $ dig +nsid ad.3fr.doubleclick.net  @209.62.177.21
> and compare with
> $ dig ad.3fr.doubleclick.net  @209.62.177.21
> 
> The 'nsid' option looks just like EDNS-PING on the wire, and causes the same
> problems.
> 

Any idea how many of them are running this or other buggy BIND versions ?

In Debian releases I've seen 9.2.1 and 9.2.3, don't know if 9.2.3 had the bug
(I haven't checked). But in that case we're talking about June 2005 or even
July 2002.

What's also interresting some guy called P. Vixie wrote the RFC for EDNS in
1999, which I presume this bug is about.

I guess we all make mistakes.

> BIND 9.2.2 has other issues too, so I hope people will upgrade quickly. In
> the meantime, I am sure we can do without the incidental ad :-)
> 
> I already tried to contact Double Click, but failed so far.
>

I guess their ad-business is not important to them. ;-)

Anyway, it's good to know it's not a bug in PowerDNS.

> 	Bert
> 



More information about the Pdns-dev mailing list