https://bugs.exim.org/show_bug.cgi?id=138
--- Comment #6 from Arkadiusz Miskiewicz <arekm@???> ---
> I have no idea if you actually get more throughput with 5000 queue-runners than
with 500
I get more because I had few hundreds (or thousands) emails being processed to
few remote MX servers which were processing these emails with speed like 1
email / few minutes. So 5000 queue runners were at least being able to deliver
other emails to other MXses. Having only 500 would mean that queue runners
would stuck for long minutes on problematic MXes and all other email would
stuck.
> deliver_queue_load_max
That would stop all delivery and be even bigger problem than a load.
Anyway all that are just workarounds.
What I'm interested in more is solving a problem where queue runners waste
system resources by processing almost empty queue (that is what causes load).
If there was some smarter scheduling then I think exim would know that there
are only few mails in queue and there is no point in waking up all 5000 runners
etc, there would be no queue scanning over and over, a lot of logging to disk
etc
--
You are receiving this mail because:
You are the QA Contact for the bug.
