As at many sites, we have people who mis-configure their mail clients by
making a mistake in their (sending) mail address.
One such case here involves Pegasus on a PC, posting via a Novell server
running Mercury. The user has set their address to something like
"username@???" (name changed to protect the guilty). When
Mercury tries to send this to one of our mail hubs (which IS
nottingham.ac.uk), I would expect Exim to reject the message because the
sender is invalid -- I have sender_verify enabled. However, the message is
instead rejected with SMTP 421,
"temporarily unable to verify sender address (try again later)
<username@???>". Mercury is supposed to retry at a configurable
interval, defaulting to about an hour, but obviously doesn't understand 421
(it logs "TCP/IP error") and retries after about 30 seconds.
What can I do to get Exim to return a 5xx permanent error, that the sending
address is, and will remain, invalid, to stop Mercury retrying?
Looking at the rejectlog, I notice that another message is being retried at
frequent, short intervals, with a sender address of
"firstname.surname.@nottingham.ac.uk" (trailing dot before the @) -- the
relevant firstname.surname alias does exist. The problem is the sending host
is another system also running Exim 1.73, so it's not just Mercury.
--
David Osborne Postmaster@???
Postmaster phone/voicemail: +44 (0)115 951 3397
University of Nottingham, UK
--
*** Exim information can be found at
http://www.exim.org/ ***
