Created an mailbox which works. Then added an alias for that, but incoming mails with that alias gets rejected.
Mailbox: ep@backup.example.com
Alias: eht@backup.example.com
Mail log:
2023-10-03T21:54:06+02:00 [1:mail2:postfix/postscreen] CONNECT from [192.168.10.10]:42090 to [192.168.10.11]:25
2023-10-03T21:54:06+02:00 [1:mail2:postfix/postscreen] PASS OLD [192.168.10.10]:42090
2023-10-03T21:54:06+02:00 [1:mail2:postfix/smtpd] connect from mail.example.com[192.168.10.10]
2023-10-03T21:54:06+02:00 [1:mail2:rspamd] #25(rspamd_proxy) <3c5420>; proxy; proxy_accept_socket: accepted milter connection from ::1 port 60662
2023-10-03T21:54:06+02:00 [1:mail2:postfix/smtpd] NOQUEUE: reject: RCPT from mail.example.com[192.168.10.10]: 550 5.1.1 <eht@backup.example.com>: Recipient address rejected: undeliverable address: lookup failed; from=<eht@example.com> to=<eht@backup.example.com> proto=ESMTP helo=<gandalf.example.com>
This evening i tested the alias once more. This time it works.
Additional i created new aliases, works as well.
I created new mailbox and a alias for it a the same time, works also.
The only thing that has changed since yesterday is that the machine was rebooted.
So if no one has an explanation or have experienced the same behavior, we might as well close this topic.
If you do a delivery attempt before the alias creation, Postfix inserts it in a cache of non-existing addresses. After a few minutes and also after a service restart the cache expires.
Just took your explanation to the test, and indeed there seems to be a cache.
So I may have used the alias address before it was created without noting, it is however very similar to the mailbox address.
Executing podman exec postfix postfix reload as the mail instance user reloads the service and there by clears the cache.
