On Fri, Jan 05, 2007 at 11:16:43AM +0100, Daniel Tiefnig wrote:
> I'm not sure though, whether this really is a bug, or one may call it a
> feature.
> Does it make sense to deliver to the 2nd MX if we know the 1st MX does
> not accept mail for the address? Or better: do we know enough about the
> setup of the target MX system to make a reasonable decision? I don't
> think so.
FWIW, the relevant sections of RFC 2821 say:
5. Address Resolution and Mail Handling
To provide reliable mail transmission, the SMTP client MUST be able to
try (and retry) each of the relevant addresses in [the list of
alternative delivery addresses because of multiple MX records] in order,
until a delivery attempt succeeds.
4.2.1 Reply Code Severities and Theory
4yz Transient Negative Completion reply
A rule of thumb to determine whether a reply fits into the 4yz or the
5yz category (see below) is that replies are 4yz if they can be
successful if repeated without any change in command form or in
properties of the sender or receiver (that is, the command is repeated
identically and the receiver does not put up a new implementation.)
This seems to leave it open to interpretation whether the sending MTA should
prefer to retry in order or to retry identically. That language isn't quite
straightforward to me, but it seems to lean more towards the former.
Ideally, we could have a separate new 45x code that would mean 'Requested
action not available: please proceed onto the other MXs'. That way there
would be no ambiguity in what the receiving MX really meant.
--
2. That which causes joy or happiness.
