On Wed, 19 Sep 2001, John Sloan wrote:
> Now we have had plenty of other messages this old on our queue which have
> now been succesfully timed out after I set this option. However this one
> is proving more stubborn. I'm curious about why I am seeing this variable
> behaviour.
Ha! It's a stupid interaction between the different options. If the
ignore_errmsg_errors_after time has passed, Exim unfreezes the message,
and sets the "ignore delivery error" flag. *Otherwise* (and that's the
mistake), if the timeout_frozen_after time is exceeded, Exim unfreezes
and sets a flag saying "timeout all recipients".
In your case, the first test is working, but the error message suffers a
*temporary* delivery failure. That doesn't count for discarding the
message. Unfortunately, it's one of the serious configuration problems
that cause the message to be frozen.
I'll fix this in Exim 4. The tests should be in the opposite order.
If you want a lash-up fix for Exim 3, remove the word "else" at the
start of line 3900 in deliver.c. I think that might sort it.
For your particular problem, you should also note the possibility of
setting
ignore_target_hosts = 127.0.0.0/8
on your lookuphost router.
--
Philip Hazel University of Cambridge Computing Service,
ph10@??? Cambridge, England. Phone: +44 1223 334714.
