Path: utzoo!utgpu!jarvis.csri.toronto.edu!cs.utexas.edu!samsung!uakari.primate.wisc.edu!ames!lll-winken!uunet!mcsun!hp4nl!ruuinf!praxis!edwin
From: oberman@rogue.llnl.gov (Oberman, Kevin)
Newsgroups: comp.protocols.tcp-ip
Subject: Re: Curious SMTP behavior
Message-ID: <50657@lll-winken.LLNL.GOV>
Date: 2 Mar 90 08:13:44 GMT
References: <48f06f20.20b6d@apollo.HP.COM>
Sender: usenet@lll-winken.LLNL.GOV
Reply-To: oberman@rogue.llnl.gov
Organization: Lawrence Livermore National Laboratory-Engineering
Lines: 27

In article <2556@ruuinf.cs.ruu.nl>, edwin@praxis.cs.ruu.nl (Edwin Kremer) writes...
>I've seen a (almost) similar behaviour on our sendmail daemon some time
>ago: it happened if sendmail was able to find the IP address for the
>domain it was resolving (in this case "apollo.com") *but* was unable to
>find MX records for that domain. By then, I concluded that it must be
>silly to define an IP address in the nameserver, and not to add MX records.
> 
>Well, I just DIGged "apollo.com" and... yep, no MX records.

Hmm. This clearly indicates a bad mailer. RFC1123 makes it clear that an MX
record should not be required for all nodes, only those providing gateway
services of come sort.

The specific flow of the lookup should be to query for an MX record first and
if any are found, try to get IP addresses for them in the order of the value of
their preferences. If 2 domains show identical preferences, the system to use
should be chosen at random. If no MX record is found, then a direct A query is
made and any respose is used as the address. Very few systems in the world have
MX records.

					R. Kevin Oberman
					Lawrence Livermore National Laboratory
					Internet: oberman@icdc.llnl.gov
   					(415) 422-6955

Disclaimer: Don't take this too seriously. I just like to improve my typing
and probably don't really know anything useful about anything.