Who does RBL checks - MailScanner or SpamAssassin?

[Logan Shaw]

On Wed, 7 Jun 2006, Res wrote:
> On Wed, 7 Jun 2006, Arthur Sherman wrote:

>> Who should do RBL checks: MailScanner or SpamAssassin?

> Neither, the MTA should do it.

That depends on your policy on what you do with spam.  Do you
tag it, or do you delete it?  At my site, we have basically
a no-delete policy; spams are tagged and passed through.
This approach gives the users more work (have to set up a
filter) but more control.

Therefore, in our case, it does little good to filter anything
at the MTA level.  We're not going to reject it at that point,
so the spam will have to go through all the stages (local
delivery, POP3 download, etc.) so it saves no work and just
adds complication.  If you are going to reject it, though,
things are different.

So in general, if you reject messages, you probably want to do
so as early as possible.  But if you tag them only, it doesn't
really matter much when you do it.

The one complication here is that with a tag-only delivery
policy, there still is one reason it'd be nice if the MTA
could know if the message is spam:  it would be helpful to
avoid sending bounce messages against undeliverable messages
if the messages are spam.  They are usually from forged e-mail
addresses anyway, so the bounces do no good and just end up
wasting resources.

   - Logan