Thanks for your support on this complex issue, Jeremy, much appreciated!
I'll try to do some more live testing/investigation if I can catch the
next mailing - I don't control this situation.
> Are all of the batch dumped in by your CRM sitting undelivered?
> Or some proportion delivered and some stuck?
> Or nearly all delivered and only a few stuck?
I could see them on mailq.
In scrolling through mainlog again, I note that lots of queue runners
were ended with "Abandon queue run: (load 12.84, max 12.00)."
While this seems likely during the deluge of receipts from the CRM (22k
in 13 minutes), after that, it seemed the load was still hovering around
too-high, causing lots of abandoned queue runs.
I increased the max load to 20 but then it just hovered around that,
abandoning queue runs without deliveries going on.
I suspect I had queue runners spawning too frequently, and the
queue_run_max set too high.
But it still seemed odd that as soon as I switched to -q not -qq
deliveries started flowing out and there were no more abandoned queue
runner jobs.
I'm currently using queue_run_max = 35, which is great in -q mode -
delivers 99% of 22k in an hour.
Is it because qq is load-intensive in phase 1? Perhaps ideally I'd have
queue_run_max set low at first (for the first pass) and then want to
increase it for delivery after.
I will try reducing queue_run_max to 5 and see how that goes with -qq.
--
## subscription configuration (requires account):
## https://lists.exim.org/mailman3/postorius/lists/exim-users.lists.exim.org/
## unsubscribe (doesn't require an account):
## [email protected]
## Exim details at http://www.exim.org/
## Please use the Wiki with this list - http://wiki.exim.org/
