Quoth Jeff Green on Sun, Mar 25, 2001:
> If I understand retries correctly, delivery of this message (there were
> approximately 12 that over the 36 hours in question arrived in the queue
> and were ultimately delivered) should have occurred every 15 minutes for
> the first two hours, then at increasing intervals for the next 16 hours,
> and finaly each 8 hours for 4 days.
But if you have queue runs every 30 minutes, it will happen only
every 30 minutes in the beginning. To make it 15, make queue
runs every 15 minutes and set retry time to 10 or so.
> My reading of the log shows that it was
> tried initially, then approximately 3 hours later when another message came
> in for the same domain, which occurred again 4 minutes later. A new message
> was received at 18:54, while the remote domain's host was still down. No
> further delivery attempts seem to have been made for the next 23 hours.
Weird. It should have waited less. The first message was
received at 09:51:14, so,if I'm not mistaken, it should have
retried at roughly 12:00, 14:00, 17:00, 21:30, ...
> I can't determine why not.
Grep the manual for delay_after_cutoff.
> The messages were finally delivered by forcing a
> queue run (exim -q) at the console.
-q is not forcing, it's suggesting. It's what run automatically
every time (hey, is the daemon still alive over there?). -qf
should have forced it (or -M).
> otherdomain.com * F,5d,24h
It's not that domain you tried to deliver to, is it?
> Personal Computer Consultant - Las Vegas, Nevada
Do they have UNIX in casinos? ;)
Vadik.
--
Time is an illusion. Lunchtime doubly so.
-- Ford Prefect
