Hello,
I've successfully set up my exim with a second configuration file to
handle our bulk deliveries in a second mail queue (as suggested by the
optimum Sheldon Hearn). The bulk setup is configured as found in the
FAQ with no queue runners automatically started, the options
queue_run_max = 1
remote_max_parallel = 1
and exim -q -C <other_conffile> processes run by a script that controls
the maximum number of queue runners in action. So far it works really
nice and everyone here is impressed.
We still have an annoying flaw: the exim -q -C <...> processes run by
the script write some warning to stderr from time to time, cron picks
those lines up and floods us with mail.
I could run those exims with stderr redirected to /dev/null, but I would
prefer to find the cause of this and have those lines go to a suitable
log file.
I could run those exims with stderr redirected to a log file, but I feel
like there should be a better solution and that that output going to
stderr is the effect os some small misconfiguration on my part.
The lines in output are usually like (private informations censored):
2001-12-17 12:25:07 exim user (uid=33) is defined only at runtime; privilege lost for -C
2001-12-17 12:25:08 => XXXX@??? R=lookuphost T=remote_smtp H=relay.sermetra.it [NNN.NNN.NNN.NNN]*
2001-12-17 12:25:08 Completed
No mention of these lines appear in the logs:
service:/var/log/exim_bulk# ls -la
total 70M
drwxr-s--- 2 www-data adm 4.0k Dec 17 06:25 ./
drwxr-xr-x 11 root root 4.0k Dec 17 06:26 ../
-rw-r----- 1 www-data adm 20M Dec 17 13:17 mainlog
-rw-r----- 1 www-data adm 11M Dec 17 06:25 mainlog.1
-rw-r----- 1 www-data adm 3.2M Dec 16 06:25 mainlog.2.gz
-rw-r----- 1 www-data adm 13M Dec 15 06:26 mainlog.3.gz
-rw-r----- 1 www-data adm 22M Dec 14 06:25 mainlog.4.gz
-rw-r----- 1 www-data adm 0 Dec 14 06:26 paniclog
-rw-r----- 1 www-data adm 276 Dec 11 14:45 paniclog.1
-rw-r----- 1 www-data adm 0 Dec 11 10:22 rejectlog
service:/var/log/exim_bulk# grep "is defined only at runtime" mainlog paniclog
service:/var/log/exim_bulk#
(The bulk exim runs as www-data because the mail is injected by PHP
scripts)
Logging is configured with:
log_file_path = /var/log/exim_bulk/%slog
log_level does not appear on the configuration file and the exim binary
is the one compiled for Debian woody, release 3.32-2)
Is there some other thing I should look at or some hint you can give
me on routing those log lines to the right direction?
Thanks in advance.
Bye, Enrico
