[Mimedefang] Thinking of implementing MimeDefang (Questions)

Jon Kenoyer jonk at otsiweb.com
Tue Jul 30 12:33:00 EDT 2002

That's exactly the set up I am using and it works great. Except we don't
user send mail to send e-mail except for outlook express remote clients.
We also have some issues because outlook express doesn't seem to be able
to filter with a client side rule based on the header information.  We
also use Mimedefang to scan for viruses and have sendmail configured to
check with the spamcop blackhole list.  End result is we went from an
ungodly amount of spam to barely a trickle.

-----Original Message-----
From: Cowles, Steve [mailto:Steve at stevecowles.com]
Sent: Tuesday, July 30, 2002 9:13 AM
To: 'mimedefang at lists.roaringpenguin.com'
Subject: [Mimedefang] Thinking of implementing MimeDefang (Questions)

Hello fellow defanger's

I'm currently contemplating implementing mimedefang. Based on what I
read so far at the defang site, this seems like the perfect solution to
in dealing with the large amounts of SPAM our e-mail server is
But before I allocate time to setup a test environment for implementing
mimedefang, I thought I would ask a few questions based on our current
e-mail architecture to see if I'm properly understanding where
fits in.

Current E-Mail architecture:

Currently, we use sendmail (8.12.5 on linux) to front-end our MS
server. Our MS Exchange server sits behind a firewall and does NOT send
receive any e-mails directly to/from the internet. Sendmail is basically
configured to relay all inbound e-mail's for our domain using the
mailertable feature. i.e. No e-mail is delivered locally on the sendmail
system. Also, our MS Exchange server runs Trend Micro's ScanMail for
Exchange for virus checking. This has all worked fine for the last
couple of

With the above in mind (and from what I have read at mimedefang), I have
come up with the following mimedefang implementation tasks/requirements
my site:
1) Must continue to use sendmail to front-end all inbound e-mail
2) Leave virus checking on Exchange Server.
3) Use MimeDefang to split e-mail header/body for SpamAssassin. 
4) Use SpamAssassin to check (flag) e-mail content for SPAM and add
appropriate header entries
5) Allow Outlook users to add a "Rule" to accept/discard e-mail based on
SpamAssassin header additions.

Item 5 is critical to me successfully implementing a SPAM solution here.
Some users here want to receive this spam garbage, but most don't. So I
would like for ALL inbound e-mail to continue to be relayed (using
sendmail/defang/spamassassin front-end) to our exchange server, but its
header modified so that Exchange/Outlook users can accept/dispose of
by custom rules. Similar to what I saw at:

Now for my questions:

1) Am I properly understanding what (where) mimedefang's role is in
processing inbound e-mail? i.e. Split e-mail into header/body for
processing, then reassemble e-mail and handoff to sendmail for final
delivery. In my case, relay to exchange server. 

2) If I implement a mimedefang/spamassassin solution, will I be able to
my existing sendmail configurations? i.e. access/mailertable entries.

FWIW: I have always wanted to implement SpamAssassin, but based on my
understanding on how to implement SpamAssassin, it required me to use
procmail. i.e. Reconfigure sendmail for local delivery. <groan!!!> If I
understand mimedefang's role, I would no longer need to use procmail.

Sorry for the long post, I'm just wanting to insure that the time I
spend on
this project is not wasted by some gotcha that I was not aware of based
our architecture/requirements. I believe I have read every link at the
mimedefang site, but feel free to point me to any/all RTFM's.

Steve Cowles
MIMEDefang mailing list
MIMEDefang at lists.roaringpenguin.com

More information about the MIMEDefang mailing list