I am currently working with a postfix cluster with 15 nodes in it. A few of the domains hosted are filtered via a different (non-spamassassin) filtering service. Most of the domains are not.
The user-accounts when sending mail use a combination of SMTP Auth and a DB table that stores the IP they authenticated with, if both = true, email is sent. Basically in order to cut down backscatter and spam/phishing hits I am looking at options to lock it down a bit more; one option is to break the cluster into two have one as normal domains and the other as ones filtered via this third party solution-very dirty considering how much scripting is in place. I have been a fan of spamassassin for awhile and would like to consider it as an option: Is there a way to configure it to block emails incoming-or-outgoing so that realms A,B and C will only allow relay based off of two conditions: A: user is authenticated and sending IP is in the table and/or B: Incoming mail comes from the pre-filtering servers? While using spamassassin normal services to filter the domains of D, E and F as normal virtual hosts (filter based on realm before the incoming spool etc.). Basically can it act act as a gateway guardian for specified realms-auto kill anything that doesnt come from an approved range for specific virtual hosts while doing normal spam filtering for the rest... if that makes sense. Any input is appreciated. Jeff -- View this message in context: http://www.nabble.com/Special-configuration-options-for-spam-assassin.-tf3991922.html#a11335502 Sent from the SpamAssassin - Users mailing list archive at Nabble.com.
