Laurent Le Moine wrote:
> I've created a Perl script who can create message files, and change the
> owner of those files to exim.
> If I run this script in /var/spool/exim/input and run the command exim
> -q, the messages are send.
>
> But if I don't run exim -q, the message stay in queue, and I notice that
> Exim check the queue every hour and send the message created by the script.
>
> Is it possible to reduce this interval?
>
> Thanks
>
From that description, it seems as if you are set to 'queue_only' and '-q1h'
When invoking the exim listener or 'supervisory' daemon, specify how often a
'cleanup' [1] queue-runner is to be dispatched with:
-q<integer><period>
Where <period> may be h for hours, m for minutes, s for seconds
EX: We run -q55s
CAVEAT: An interval that short is a bit on the aggressive side. OTOH, we don't
need IRC
;-)
-q2m or -q5m
might be a better 'typical' choice.
In the typical *BSD installation that appears in /etc/rc.conf as part of;
exim_enable="YES"
exim_flags="-q55s"
Or could be in the ~/etc/local/rc.d/<exim calling script>
Linux, not my area of expertise, might put that "-q2m" into part of an init script.
HTH
Bill
[1] IF you are NOT running 'queue_only' each message traversing the MTA
kicks-off a queue-runner of its own, and the '-q' setting is not overly critical.
If that 'coupled' delivery is balked, however, the queue-runner fired by timeout
of the command-tail setting (above) insures that whatever is hanging around the
queue is retried if/as/when its retry time has arrived or passed.
Accordingly, the interval should not be any longer than your shortest retry
period - generally between 2 and 15 minutes, commonly around 5 minutes.
If you ARE running 'queue_only' as you seem to be (why so?) then it is more than
just a clean-up / retry tool - it is essential for ANY delivery, balked or not.
