[COLUG] effectiveness of greylisting.

[Chris Clonch]

On Saturday 24 November 2007 8:35:18 Duane wrote:
> I'm split over this, yes there is benefit in greylisting as soon as
> possible, however there is benefit in greylisting selectively only if
> the mail seems even remotely spammy also.

I swear by it.  Been using postgrey (with postfix) for ~2 years and my spam 
dropped to 0.  Now it has slowly been creepping back but that is do to 
spammers hitting my backup MX.  It through a hosting service and not 
controlled by me.  Since they run an RFC complaint server, they honor the 
resend, and the spam eventually gets through.

> In the end it comes down to personal preference, resources at hand etc
> etc etc and at this stage I'm wanting to trail greylisting if
> amavis/spamassassin says the message is > 0 score.

I would filter through greylisting first as it should avoid the resource hit 
of amavis/spamassassin processing all messages first (at least you can weed 
out the easy non-complaint MTA sent messages).  The nice piece of postgrey is 
the auto-whitelisting as Rob mentioned.  Just thinking about it, I wounder if 
postgrey can add a tag to the header to reduce needed resources on subsequent 
spam processing.

> I do on the other hand enforce strict RFC checking of connections and
> reverse hostname lookups and such, it's amazing how much spam this stops
> dead in its tracks. I've had very few instances of legit servers being
> incorrectly configured, nothing a quick chat hasn't fixed anyway,
> although the rejection message usually spells it out fairly clearly. :)

This probably follows the same reasoning for greylisting: non-RFC complaint 
servers and such.

-Chris