query-source for multiple interfaces

Sam Wilson Sam.Wilson at ed.ac.uk
Fri May 19 09:48:35 UTC 2006


In article <e4iujv$242$1 at sf1.isc.org>,
 Mark Andrews <Mark_Andrews at isc.org> wrote:

>  Well he actually said "multicast" which a assummed was his
>  way of saying "*".

AARGGH!  Cross pollution from a different area of the brain/desk.  I  
meant anycast.

>  named won't send from a real multicast address (class D) as
>  multicast needs code support which doesn't exist.
> 
>  In general there is no way to select a particular address to
>  sending to a set of addresses.  If someone feels like adding
>  it we will look at the patches.  Normally the kernel selects
>  a appropriate address (based on outgoing interface, etc.) for
>  multihomed boxes.

Yeah - that's fine and Barry has convinced me that it's not a problem.

>  If he actually ment anycast, then you only need to ensure that
>  there us a unicast address configured first or the routing
>  doesn't send the packet out that interface.  It will still receive
>  packets to the unicast address.  Again this is just using the
>  kernel's address selection rules.

Since the anycast address is an alias on the loopback interface that 
should be fine.

>  Replies to queries addresses to the unicast address will be replied
>  to using the unicast address as the source address.
> 
>  Mark
> 
>  I would expect something like this would work:
> 
>     query-source address 1.2.3.4 port 53 { !192.168/16; any; };
>     query-source address 192.168.3.4 port 53 { 192.168/16; };

Wha...?  That syntax isn't mentioned in any copy of the ARM that I can 
find.  I see that the BIND 9.4 page at ISC says, under new features:

  Accept 'notify-source' style syntax for query-source.

and under Changes there is:

1597. [func]      Allow notify-source and query-source to be specified
         on a per server basis similar to transfer-source.
         [RT #6496]

But the ARM gives the same syntax for query-source, notify-source and 
transfer-source, namely without the "{...}".

But it's OK because I don't need to do that! :-)

Sam Wilson                             one of hostmaster at ed.ac.uk
Infrastructure Services Division
Computing Services, The University of Edinburgh
Edinburgh, Scotland, UK



More information about the bind-users mailing list