On Wed, 19 Jun 2002, Tom Kistner wrote:
> I assumed that "-Mc" was the right thing to do, since exim calls itself
> in that manner to deliver messages. Now, it seems that the retry
> database is not correctly updated in this scenario.
>
> To Philip: are there any caveats when calling "-Mc" externally ?
There shouldn't be. The only difference between -Mc and -M is that the
former respects retry times, hold_domains, etc. whereas the latter does
not (and it also automatically thaws).
If a message can't be delivered, the retry database should get updated
whichever you use, because the code that runs at the end of a delivery
attempt is the same in all cases.
> BTW changing it to "exim -q <ID> <ID>" does not seem to help ...
I wouldn't expect it to. The only difference between -q and -Mc is that
the former counts as a "queue run".
If there's further need to investigate this problem (I haven't paid much
attention so far :-) the thing to do is to look at a deferred delivery
and see what is in the retry database for its domains. Also check its
msglog to see what's going on, and of course the main log to see if the
queue runners are processing it as they should.
--
Philip Hazel University of Cambridge Computing Service,
ph10@??? Cambridge, England. Phone: +44 1223 334714.
