I asked a short time ago:
> My question is: why isn't the 451 failure causing Exim to re-try the
> message at the times I've specified? If this ought to be obvious from
> the documentation, sorry, there is a network problem between here and
> the exim.org web site so I am having difficulty reading the
> documentation at the moment.
Ah, sorry -- I've just re-discovered Philip's message of 19th October
concerning 4xx errors. I'll add my two euros' worth and say that I
agree with Philip that basing the re-trying on the combination of
message id and remote host when the 4xx is returned after the MAIL
FROM or RCPT TO is an excellent one -- in fact I think I had assumed
that it was already implemented. I am definitely voting for it on the
wish-list.
For what it's worth, I finally got rid of the messages by setting
ignore-errmsg-errors in the general configuration and serialize_hosts
for the particular destination in the remote_smtp transport. The
serialize_hosts stopped us from hammering the destination machine so
hard, so that it was able to receive each message and bounce it with a
5xx error (since of course the fake reply address was non-existant)
and we then threw the bounced messages away instead of freezing them
and then auto-thawing them later.
Gyan.
--
*** Exim information can be found at
http://www.exim.org/ ***
