> When a spammer forges an existent mail address, and you're getting tons af > bounces (out of office, notice that the mail has been blocked because it's > spam, callouts, Abwesenheitsnotiz, autoresponders, ndr, confirmation > requests, all kind of challenge response) it's not so easy to block > backscatter.
Backscatter is NDR messages. Vacation (those are spam, especially if implemented improperly, but not backscatter) and other auto-responder/confirmation type messages are NOT backscatter, but may be considered spam by some people. > And in the http://www.postfix.org/BACKSCATTER_README.html you find some of > the best suggestions. Yes - but those recommendations are only directly relevant for a direct-internet-facing postfix box. By directly I mean for your postfix config - the philosophy is sound regardless, but if you have ASSP in front of your postfix box, it is ASSP that will be generating the 550 reject status codes (again, these are not EMAIL messages), not postfix. More importantly, you want to make sure that postfix does NOT generate any NDRs under ANY circumstances, if it has ASSP in front of it. I'll say it again in different words... If you pass email through ASSP to postfix and then let postfix generate the NDR, then YOU ARE GENERATING BACKSCATTER, because ASSP has ALREADY ACCEPTED THE MESSAGE FOR DELIVERY. You can NOT SMTP-REJECT a message after it has been accepted for delivery - if you do, you are generating backscatter. > I just don't care when when backscatter is directed to nonexistent > addresses, that's not a problem. Problem is when a user start getting > thousand bounces per hour, with all kind of exotic messages. And I'm trying > to keep his mailbox readable. So, reject/blackhole NDR/backscatter messages that do not originate from your own MTA, as that is the only one that will generate legitimate NDRs. -- Best regards, Charles ------------------------------------------------------------------------- This SF.net email is sponsored by: Microsoft Defy all challenges. Microsoft(R) Visual Studio 2005. http://clk.atdmt.com/MRT/go/vse0120000070mrt/direct/01/ _______________________________________________ Assp-user mailing list [email protected] https://lists.sourceforge.net/lists/listinfo/assp-user
