John C Klensin wrote:
--On Monday, 23 April, 2007 09:03 -0700
ned+ietf-smtp(_at_)mrochek(_dot_)com wrote:
[David Skoll]
If an SMTP client receives a reply code that is not defined
by this RFC or a negotiated extension, it SHOULD send a RSET
or QUIT command, and SHOULD treat the unknown reply code as a
5xx code.
[Ned]
I'm with you up to the 5yz code. I think 4yz would be more
appropriate - that way you don't bounce messages unnecessarily
when dealing with a broken server.
[John]
Hmm. Do you want to requeue it and keep trying for four or five
days on the theory that the server might spontaneously fix
itself? Or would it be better to bounce the puppy and hope that
someone will complain? I think I can argue that one either
way.
I prefer 5yz. It makes the problem immediately obvious; there's
nothing worse than finding out days later that important e-mail didn't
get through because of a broken server. As a general principle, I
think anything that doesn't "make sense" should be treated as 5yz. I
have a lot of experience from our support line that mail
administrators often ignore 4yz failures in their logs, whereas they
react pretty quickly to complaints about bounces!
However, it's a SHOULD, so a client could treat it as a 4yz if it
really wants. Actually, the more I think about it, the more it might
make sense to treat it as a 4yz if there are any lower-priority MX
hosts left to try, and a 5yz if not. But putting that in an RFC is
kind of ugly; I wouldn't know how to word it nicely.
Regards,
David.