Cary, > I upgraded from 2.6 spammassassin to the latest I have no idea and I am too > tired to look what I need to know is I am going to switch the bayes to use > mysql but I cant backup my old ones as the instructions I found used a > sa-learn --backup which is not available in the sa-learn in 2.6, is it the > same as sa-learn --dump?
Please ask on the SA list ([email protected]), you'll get a more qualified answer there. Or you may start from a fresh database, lower bayes scores for few days until it auto-learns to distinguish ham from spam, but make sure you have a good set of other rules, like most network tests (Razor, DCC, RBL), the usual SARE rules (http://www.rulesemporium.com/) and updated SA 3.1.x rules by the new SA utility sa-update. This will provide good initial Bayes training, and cover for the few-days period while it learns, then you can bring Bayes scores back to default values. One can think of Bayes as a big flywheel, it builds its momentum from other rules, then gives it back when other rules develop a temporary lapse. Mark _______________________________________________ AMaViS-user mailing list [email protected] https://lists.sourceforge.net/lists/listinfo/amavis-user AMaViS-FAQ:http://www.amavis.org/amavis-faq.php3 AMaViS-HowTos:http://www.amavis.org/howto/
