multiple mailscanner instances?

Julian Field mailscanner at ecs.soton.ac.uk
Mon Sep 16 20:00:21 IST 2002


At 01:44 14/09/2002, you wrote:
>Julian could provide the best analysis, but a quick read of the code
>said to me that it scans the input queue and identifies "Max
>Safe|Unsafe" sized batch of messages to process. The header data of
>those are written into MailScanners incoming directory
>(xxxxxxxxxx.header) and an internal list is constructed. Then it
>processes each message in that batch. The original copy of each message
>remains in the input queue until processing of that batch is complete.
>
>So running more than one instance on the same input queue and with the
>same work dirs isn't going to work, if I read the code correctly. At the
>very least you'd get duplicate delivery of at least some messages.

Exactly right.

>Distributing the input queue to multiple processing queues is cheap,
>because it's just a "mv" of the qf/df pair. If you are interested I can
>clean up the queue mgmt code and document how I do it on a server that
>handles some 150-180K messages a day.

Waiting for V4 is much easier :-)
--
Julian Field                Teaching Systems Manager
jkf at ecs.soton.ac.uk         Dept. of Electronics & Computer Science
Tel. 023 8059 2817          University of Southampton
                             Southampton SO17 1BJ



More information about the MailScanner mailing list