[Mailmunge] Temp fail due to unable to start worker.

[Dianne Skoll]

On Mon, 09 Aug 2021 19:50:45 -0400
postfix--- via Mailmunge <mailmunge at lists.mailmunge.org> wrote:

> I had an incoming email temp fail and the logs said

> Where did i short change myself? Is it the MX_MIN_WORKER_DELAY

Yes.

> and if i set that to 0 will it spawn new workers as fast as needed
> during burst?

Yes.

I recommend MX_MIN_WORKER_DELAY be kept at 0 on all but the very busiest
servers.

> This incident happened when postfix was handling 10 concurrent
> connections from the same client. Im also wondering if workers were
> busy why didn't the connection get queued instead of temp failed?

Ah... the queueing logic only kicks in if all workers are busy.  It
doesn't kick in if a worker is theoretically free, but could not be
started.  That's arguably a bug (although it is how the docs specify
that it works), but it's a rare edge-case, especially if
MX_MIN_WORKER_DELAY is set to zero, so it probably won't be fixed.

Regards,

Dianne.