Fixes for possible remote email problems

I have RT3.0.3 running happily with RedHat 8.0 and PosgreSQL 7.2.
However, it took a while - not because of RT3, but for other
aspects of my setup. I thought I’d report the problems and
the fixes I found just in case anyone else has a similar setup.

Setup: One machine IP address 1.2.3.4 hosting multiple virtual
domains. One of which: test.Y.Z, is reserved for RT
use. RT3 is configured to use Geert van der Pleog’s
procmail setup so there is no need to add queue names to
/etc/aliases (all email to to test.Y.Z gets processed by
procmail and piped to rt-mailgate).

Problem 1: Email to queues from outside machines was simply
vanishing.

Reason: The domain name test.Y.Z was define in DNS
as a CNAME (DNS term) record pointing to Y.Z, so
the remote MTA was converting queue@test.Y.Z
to queue@Y.Z before sending the email. (This
took a while to find because queue@Y.Z was a valid
user, hence the vanishing instead of bouncing).

Solution: Change the CNAME record to an A (DNS term
again) record pointing directly to 1.2.3.4. The
problem went away as soon as the DNS tables were
propogated.

Problem 2: Outgoing RT3 messages were having their Repy-To:
addresses converted from ‘queue@test.Y.Z’ to
‘queue@Y.Z’ so replies to comments and responses
also vanished.

Reason: Sendmail was configured to masquerade outgoing
email to Y.Z.

Solution: Add MASQUERADE_EXCEPTION(`test.Y.Z’)dnl to
/etc/mail/sendmail.mc, remake sendmail.cf and
restart sendmail.

I hope someone finds this useful.
Steve

Steve Wampler {sbw@tapestry.tucson.az.us}
The gods that smiled upon your birth are laughing now. – fortune cookie