On Wed, Dec 16, 2009 at 01:08:54AM -0600, Adam Thompson wrote: I'm not sure MTA is the right question.
I leave most of the anti-spam heavy lifting to OpenBSD's spamd with greylisting, greytrapping, a couple explicit blacklists (uatraps, nixspam, china, and korea), and a whitelist for the handful of places I care about that can't handle greylisting (Google, fedex...). As my dad and I still post quite a bit to USENET there's also a script that looks for GREY entries with To: addresses matching USENET message-IDs to add to the greytrap list. These days spam that gets past spamd is pretty much limited to hotmail, hosts with compromised web form by mail cgi-bins, and the occasional new foreign ISP. After that it gets fed through an MTA with the mailscanner milter, which seems to be a minimally painful way to integrate with spamassassin, anti-virus, windows executable detection and quarantine, and flagging of nigerian bank scams and such. I currently use postfix for the MTA, but have also recently configured other MTAs because they were less of a PITA to configure to different virtualhost/virtualuser backends for IMAP. -- Chris Dukes
