On 09/04/15 14:26, Rob Gunther wrote:
> In our normal day-to-day we have maybe 2,500 messages in the exim queue at
> any time. We retry delivery every 15 minutes, slowing down the retry as
> messages gets older.
>
> Had a bit of a routing issue, where Exim was accepting mail but could not
> deliver to the next server (not a mailbox).
>
> The queue grew on one server to just over 200,000 messages ( 14 GIGS worth!
> ).
>
> We stopped sending new mail to the server, so it could deal with what it
> had.
>
> I assumed it would blast through those messages very quick.
>
> It didn't move quick, it seemed to crawl.
Exim does have a rep. for behaving that way. The writer (I think)
commented that it was traditionally intended for environments
where the queue could be kept small.
There is some built-in support for spreading the queue over
several directories. See split_spool_directory.
Some sites implement things like: once an item is old-enough
in the queue, move it to a different queue directory, which
is serviced by a separate exim instance.
See also
https://github.com/Exim/exim/wiki/Performance
--
Cheers,
Jeremy
