named, RedHat 6.2, and multicast interaction problem

[Antonio Querubin]

On a Redhat Linux 6.2 system, named (BIND 8.2.2 P5-9) syslogs numerous
messages of the form:

"refused query on non-query socket from [x.x.x.x].1024"

where x.x.x.x is the address of a host from which it is receiving a
multicast UDP stream from.  These syslog messages stop if there are no
more processes that have 'joined' the multicast group.  The multicast
group address 224.0.1.143 and the destination port is 1024.

A BSDI 4.x system does not exhibit this syslog problem when it's receiving
the same multicast stream.  So it seems that this is peculiar to Redhat
6.2 and the version of BIND that comes with it.  Is this a known problem
and a possible fix?

-- 
Antonio Querubin
tony at lava.net