Something else strange, i have been trying to replicate this situation
using a dummy server to send my server a message with From: and To:
using the same invalid address. This time i got two reject messages in
the logs:
NOQUEUE: reject: RCPT from example2.com[IP]: 550 5.1.0
<u...@example.com>: Sender address rejected: User unknown in virtual
mailbox table; from=<u...@example.com> to=<u...@example.com> proto=ESMTP
helo=<example2.com>
NOQUEUE: milter-reject: RCPT from example2.com[IP]: 550 5.5.1 Address
u...@example.com does not exist; from=<> to=<u...@example.com>
proto=ESMTP helo=<example2.com>
So not only did the first reject come from the postfix default when i
have smtpd_reject_unlisted_recipient = no, but it did a second check?
Shouldn't it have been done and over on the first reject? Now the second
reject came from the milter and had from=<> which i assume is when its
trying to deliver a postmaster message like a rejected back splatter
email. But why would postfix be trying to send a back splatter email
when it never accepted a message and already rejected the rcpt command?
And yes both of these rejected log lines are part of the same email and
are one right after the other. The entire transaction was only 4 log
lines, the connect, those two rejects, and then the disconnect message.
Im confused by all of this, can you shed light on what is happening?
