[Pdns-users] Question about Multiple CNAMES

Clint Martin clintm at laplink.com
Mon Oct 11 14:19:24 UTC 2004

Clint Martin
Network Administrator
LapLink Software Inc.

> -----Original Message-----
> From: Lorens Kockum [mailto:lk-sender-02 at lorens.com] On Behalf Of
Lorens Kockum
> Sent: Saturday, October 09, 2004 2:57 AM
> To: Clint Martin
> Cc: pdns-users at mailman.powerdns.com
> Subject: Re: [Pdns-users] Question about Multiple CNAMES
> On Fri, Oct 08, 2004 at 01:39:34PM -0700, Clint Martin wrote:
>> I can accomplish my goals this way.. This is how I have done it in
>> past..  I don't mind having to change my setup if this one just won't

>> work, or is fundamentally flawed. ..
>RFC 1034 section 3.6.2 says "A CNAME RR identifies its owner name as an
alias, and specifies the corresponding canonical >name in the RDATA
section of the RR.  If a CNAME RR is present at a node, no other data
should be present".
>AFAICS nothing in RFCs 1034 and 1035 address multiple cnames, however a
few seconds of googling turns up
>where it is noted that bind 8 will throw an error on your
configuration, but that setting bind option "multiple-cnames"
>will allow it, and that A records in that configuration will be
returned in a round-robin fashion.
>So, your configuration is in a very grey area :-) No idea if pdns
handles your configuration with round-robin.

Thank you very much.. I would like to see this situation handled by
either all of the relevant A records being returned, or one being
returned in a Round-Robin Fashion.. Who knows tho.. Perhaps I'll take
some time to dig through the source.. I've been meaning to look at it
for a while now anyway..

>> I was trying to keep the references to specific IP addresses limited 
>> to a bare few number of records.. This would allow me to modify the 
>> domain far easier and simpler than having many, many A records..
>update records set content = '' where content = '' ;

This works great.. As long as the only instances of the IP are
ones that I want to change..  Unfortunately, this is not always the


>#include <std_disclaim.h>                          Lorens Kockum

More information about the Pdns-users mailing list