For some time (since Oct 2022) I’ve been receiving 2 copies of the weekly usage report. I put it down to “doesn’t really matter” and “it will get fixed sometime”. There are a couple of old forum posts about this, where the problem just went away, but no cause or fix found. Is anyone else seeing this, or found a fix?
I’ve looked through the code and these’s nothing obvious (!) in the code, or in my config. I’ve only got one admin account, and the email is only sent to “administrator@machineprimarydomain” anyway. The emails are always identical and have the same message id. I don’t have any alias that would do this. And all the other status messages are normal/single. Weird
Oddly, I just saw this duplicate “Mail-in-a-Box Usage Report” for the first time. The Message-ID is identical, but I diffed the header properties and noted this difference:
Interesting … I’ve done some more hunting around. It looks like I’ve got an amusing timing issue
I end up with two emails having the same initial Message-Id but with different lmtp ids. (Which, for extra points, seems to confuse Thunderbird because it sometimes shows me the same envelope for both emails, unless I really go hunting )
I’ve looked through the logs. My system generates the “Mail-in-a-Box Usage Report” email at 1:05:08 and the message is it’s making it’s way through the email pipeline when … at 1:05:11, everything is clobbered for the backup.
Then system email daemons are bought back up a minute later, they find the not-completely-handled email, and redeliver it.
It’s really impressive that this kind of timing hassle is happening quite reliably!