[Pdns-users] pdns recursor together with bind9 (IPv6 issue?)

[Marco Davids]

Hi,

I ran into a problem while trying to run both bind9 and pdns_recursur  
on the same box.

Bind9 can be configured to listen on certain IPv4 addresses, but for  
IPv6 the possible choices are either 'any'  or 'none'. It seems this  
is the root-cause of my problem, but I do not quite understand  
precisely why.

Here is my situation:

pdns_recursor configured for local-address 192.168.1.1, interface eth1.
bind9 configured to listen on 192.168.2.2, interface eth0  (and  
obviously on all IPv6 addresses, including the one on eth1)

When I start bind9 first and then pdns_recursor, I get:

Exception: Resolver binding to server socket on port 53 for  
192.168.1.1: Address already in use

(netstat clearly shows that bind9 is *not* in the way on that ipv4  
address though)

When I configure bind9 not to use IPv6 at all, both of the  
applications run happily together.

When I start pdns_recursor first and then bind9 (with IPv6 enabled)  
it complains in a similar way.

Can this be fixed, or is this expected behaviour and am I missing  
something?

OS:				Debian Linux
Kernel:			2.6.8-3-k7
Bind9:			9.2.4-1sarge1
pdns_recursor:	 3.1.3 static .deb from powerDNS website

Thanks!

--
Marco




-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mailman.powerdns.com/pipermail/pdns-users/attachments/20061008/a7e9c4b9/attachment.html>