1.) No spam filtering solution will catch all spam. Play with your SA rules or enable some blacklists. 2.) You should configure DSPAM as a plugin into SA. 3.) SA is better tested then DSPAM, imo. Run dspam as procmail script and see how well it works, running after mimedefang/sa then slowly move it into production usage after a few weeks at minimum. You can also run sa as a procmail script after mimedefang, that allows it to generate individual bayes dbase/awl per user. I use this method on one of my systems. Are we talking about just your inbox, or a mail server with lots of users?
Andrew S. Johnson said: > No matter how much I keep training the SpamAssassin Bayesian > filter, it still leaks spam into my inbox. As reported on /. DSPAM > is supposed to be a better algorithm. I could just do a rip and > replace, but I'd like to keep the MimeDefang AV scanning that I > have with ClamAV. Has anyone set this up before I get started > on my own, or should I just reconfigure sendmail to use ClamAV > and DSPAM just through milters? My theory is that if I keep the > existing MimeDefang setup I could easily switch between SA and > DSPAM until I get comfortable with DSPAM. -- Luke Computer Science System Administrator Security Administrator,College of Engineering Montana State University-Bozeman,Montana _______________________________________________ Visit http://www.mimedefang.org and http://www.canit.ca MIMEDefang mailing list [EMAIL PROTECTED] http://lists.roaringpenguin.com/mailman/listinfo/mimedefang
