On Tue, 26 Jun 2001, Mark Baker wrote:
> When exim fails to deliver a message to postmaster, it probably
> shouldn't report the failure by sending another message to postmaster...
Without more information as to the configuration that was in use, and
preferable some -d9 output of a failing case, I can't comment on this.
(Certainly a "normal" message from X to postmaster should send a bounce
message to X, so there is something unusual going on.)
> (The loop detection needs to work even when postmaster ultimately
> expands to multiple addresses and only one of them fails, which is what
> was happening in my case.)
OK, that's a tiny snippet more information, but not enough. Who sent
mail to postmaster?
Note: If this is what I think it might be, I don't know of any good
solution, and it's nothing to do with the postmaster address. Another
scenario that is often seen is:
1. User creates a .forward with a bad address in it.
2. User tests the .forward by sending a message *from the same account*.
3. Message fails, generates a bounce to sender.
4. Bounce fails; gets frozen.
5. User gets impatient; sends another message ...
What can Exim do? There are some options for bypassing .forwards that
fail (I forget the details and I'm at home at the moment). The only
possibilities are freezing (as now) or throwing the message away. (Or
bothering the postmaster, but on busy systems that probably isn't a good
idea.)
--
Philip Hazel University of Cambridge Computing Service,
ph10@??? Cambridge, England. Phone: +44 1223 334714.
