[Pdns-users] PDNS devel and weird IPv6 log entries

[Detlef Peeters]

On Tue, 14 Jun 2011 15:18:09 +0200, Detlef Peeters wrote:
> On Tue, 14 Jun 2011 11:55:14 +0200, bert hubert wrote:
>> On Tue, Jun 14, 2011 at 11:21:45AM +0200, Detlef Peeters wrote:
>>> >Or to any of the slaves of depee.org ?
>>>
>>> Yes, IPv6 is enabled, and PowerDNS ist listening to it. The mater
>>> server for the domain depee.org has the IPv6 address:
>>> "2001:4d88:ffff:ffff:d0:b723:6daf:2".
>>>
>>> With PowerDNS 2.9.22 and the same configuration I haven't seen this
>>> log entries.
>>
>> This is probably because PowerDNS 2.9.22 would not notify IPv6 hosts 
>> ;-) The
>> error is very weird, could you tcpdump to see if the correct IP 
>> addresses
>> are actually on the wire?
>
> Yes, the correct IPv6 address is on the interface. This log entry is
> not on each zone update. So it's not so easy to debug.
>
> I can also see the following new log entry:
>
> Jun 14 15:05:57 eka pdns[71532]: No question section in packet from
> 2a00:dd0:0:6::17, rcode=5
> Jun 14 15:05:57 eka pdns[71532]: Unable to parse SOA notification
> answer from 2a00:dd0:0:6::17
>
> The Server with the IP 2a00:dd0:0:6::17 has as DNS Server NSD 3.2.7
> with IPv6 enabled and the zone is correct updated on the secondary.

O.K. I'll answering myself. ;) I've done some checks and the problem 
was with the ip6tables under CentOS 5. I've done som changes and now it' 
working.

I've also done the check with the current 
pdns-static-3.0rc2.20110614.2216-1.i386.rpm and the zone transfer and 
update is working via IPv6.

regards,

Detlef