> I have received questions regarding messages that appear to either be
coming
> from or sent to ig1/ig2.intouchmi.com (my Imgate boxes) in the From/To
> headers. Some are legit messages, some like the sample below are
definately
> spam.
This is why I have the following:
smtpd_recipient_restrictions =
reject_unauth_pipelining,
reject_non_fqdn_recipient,
reject_non_fqdn_sender,
That way the non-fqdn stuff is killed early, and the append_at_myorigin is
a non-issue.
> I don't see a useful reason for Postfix touching the message header -
this
> isn't the info thats used to route the messages anyway(?)
It has to touch the header. All mail servers do to enter the added lines.
There are plenty of other reasons, such as virtual mapping, canonical
changes, so on and so forth.
One of the most typical reasons to do the append_at_myorigin is for local
users sending mail. That way they show up as @whatever rather than a
non-fqdn.
In a relay only anti-spam server, like the IMGate, many of the reasons for
a MTA to fiddle with addresses are gone. And that is why you wonder as to
the purpose. But never forget this is intended as an MTA for a POSIX OS
that may have local users, and other issues.
I hope that helps.
--Eric
