<domain>
The default is no. The line specifies how exactly the interruption allowed by the neighboring allow-priority-interruption happens.
With the default the interruption is postponed until the connection frees up — that is, the current message is finished, and only then is the connection handed over.
With the value yes the interruption happens immediately — in the middle of a transfer.
A separate caveat: the line concerns only SMTP queues.
So the time gained for the urgent queue is paid for with a retransmission in the slow one — and with a reputation risk at the receiver that saw the drop.
The vendor chose the default deliberately: waiting for the connection to free up is almost always cheaper.
The Hub does not set priorities at all, so there is nothing to interrupt: the line will not show up at any value.
What to set. The default is almost always right:
| Value | When |
|---|---|
| no (default) | the usual case: waiting for the connection to free up is cheaper than sending the message again |
| yes | urgency outweighs the cost of retransmission — a rare case, usually transactional mail in a separate stream |
The second is worth weighing twice: a drop in the middle of a transfer is visible to the receiver, and the message will be sent again anyway.
# an urgent queue does not wait for a connection to free up — at the cost of a retransmission <domain transactional.example> queue-priority 90 allow-priority-interruption-during-transfer true</domain>
what people get wrong and how it ends2 mistakes▶
§ 3.3.10.13
§ 3.3.10.13
1 relation
grep -n 'allow-priority-interruption' /etc/pmta/config /etc/pmta/virtualhost.txt grep -i 'connection.*reset\|abort' /var/log/pmta/pmta.log | tail -10
3 articles
◇What the node tells about itself and through which channelsSeven observation surfaces: the web monitor, the console in four formats, the log, accounting, accounting search, metrics and export to external analytics.◇The watchdog that restarts the engineKeeps an eye on the engine and brings it back up after a crash by itself, sending a crash report to the postmaster address.◇What makes a virtual machine riskyRunning on a virtual machine is possible, but the manual does not recommend it without sufficient hardware and properly tuned hypervisor settings: under load the node may stop responding.