Strange slave DNS errors

[Stefan Puiu]

Search the list archives (http://marc.theaimsgroup.com/?l=3Dbind-users)
for "bind 9.2,1 internal_send", you'd be surprised to see how many
times this has been mentioned before. There are a few solutions
provided (patch, upgrade), you just have to see which suits your
needs.

On 8/2/05, Kirk Hoganson <kirk2 at lenderlab.com> wrote:
> I am having a rather strange issue with one of my DNS slave servers.  We
> have 3 DNS servers:
> Master: BIND 9.2.1 (Debian)
> Slave1: BIND 9.2.1 (Debian)
> Slave2: BIND 9.2.2-P1 (Gentoo)
>=20
> The first slave fills the logs with the following error:
> Aug  2 00:23:36 ns2.domain.com named[1972]: socket.c:1100: unexpected err=
or:
> Aug  2 00:23:36 ns2.domain.com named[1972]: internal_send:
> 66.7.6.158#53: Invalid argument
> Aug  2 00:23:44 ns2.domain.com named[1972]: socket.c:1100: unexpected err=
or:
> Aug  2 00:23:44 ns2.domain.com named[1972]: internal_send:
> 66.7.6.158#53: Invalid argument
>=20
> (That is not the actual IP, though the actual IP is valid).
>=20
> I have can find no reference to this error when using a valid IP.
> Further, the service will hang when I attempt to shut it down, and I
> must kill the processes to close it.
>=20
> The server will respond to outside DNS requests (ie. dig), but any local
> requests fail.
>=20
> The other two systems do not have this problem, and seem to function
> correctly.
>=20
> This is causing me serious grief, can anyone give me some information
> that might get me pointed in the right direction?
>=20
> Kirk
>=20
>=20
>