Ignoring MX records - how common

Mark_Andrews at isc.org Mark_Andrews at isc.org
Mon Feb 10 20:37:17 UTC 2003


> Simon Waters <Simon at wretched.demon.co.uk> wrote:
> 
> >Having changed the configuration of my MTA recently I noted that
> >a handful of genuine e-mails are delivered (and thus now lost)
> >to the A record for the domain wretched.demon.co.uk
> >
> >The domain has MX records, and my reading of the RFC's (and
> >Eric's if his book on sendmail is to be believed) says these
> >should always be honoured if they exist, which agrees with
> >commonsense.
> >
> >Anyone have a feeling for how big this "issue" is?
> 
> I hadn't noticed it before, but a moment I ago just saw in my log that
>    bay1-f189.bay1.hotmail.com[65.54.245.189]
> had attempted to deliver mail to the A record machine, even though there
> are two MX records.  I know that the secondary MX was running at the
> time, and I have no particular reason to believe that the primary MX was
> down.
> 
> I have of course verified that the IP address and the name maps to each
> other both ways.  The sender address seemed plausible as a genuine
> sender - I don't think this was spam.
> 
> It is a problem if hotmail does this in general.
> 
> -- 
> Jesper Dybdal, Denmark.
> http://www.dybdal.dk (in Danish).
	
	It will be SPAM unless one of your nameservers is returning
	NODATA to MX requests.  You have to have a negative answer
	to the MX query before you make the A request.  Nameserver
	failure is *not* enough to cause fallback to A records.

	Mark
--
Mark Andrews, Internet Software Consortium
1 Seymour St., Dundas Valley, NSW 2117, Australia
PHONE: +61 2 9871 4742                 INTERNET: Mark.Andrews at isc.org


More information about the bind-users mailing list