[ Apologies if anyone has seen this before. ]
I'm seeing problems with:
Exim 1.82.
Linux/Intel 2.0.33
DB 1.85 with various RedHat 4.2 bugfix patches
However, this problem occured during 1.7x and with various earlier Linux
and DB sets of software.
Exim is configured to treat certain domains as local, run them through
various alias lists, then relay them on to a second mail server.
After a mail server has been down for even a very brief time (and
sometimes when a server hasn't been down at all) and after a long period
of uptime, we're seeing some emails delivered successfully but many
bounced with the error message "all hosts have been failing for a long
time and were last tried after this message arrived".
1998-02-09 12:32:29 Start queue run: pid=8089
1998-02-09 12:32:29 0y16sK-0002WG-00 => user_a@???
D=relay_local T=relay_smtp H=mail.cust.co.uk.localnet [10.1.1.1]
1998-02-09 12:32:29 0y16sK-0002WG-00 Completed
1998-02-09 12:32:30 0y1sN5-0001x0-00 ** user_b@???
D=relay_local T=relay_smtp: all hosts have been failing for a long time and
were last tried after this message arrived
1998-02-09 12:32:31 0y1sMB-0001xe-00 => user_c@???
D=relay_local T=relay_smtp H=mail.cust.co.uk.localnet. [10.1.1.1]
1998-02-09 12:32:31 0y1sMB-0001xe-00 Completed
1998-02-09 12:32:32 0y1sKK-0001p4-00 ** user_d@???
D=relay_local T=relay_smtp: all hosts have been failing for a long time and
were last tried after this message arrived
In this instance, every second instance is failing but this varies. Exim
doesn't seem to be retrying the messages before bouncing them despite them
only having spent a couple of hours in the queue, the mail server only
having been down for about a day and the "no_delay_after_cutoff" option.
Relevant director:
relay_local:
driver = smartuser,
transport = relay_smtp;
Relevant transport:
relay_smtp:
driver = smtp;
hosts = mail.cust.co.uk.,
no_delay_after_cutoff,
final_timeout = 30m
Any ideas what's wrong? I have a snaphost of the db directory while this
is occuring if it is likely to help.
Steve Clarke
--
*** Exim information can be found at
http://www.exim.org/ ***
