>I'm running SA 3.4 and am getting several messages like both of these, headers below, that are getting through. I've run sa-learn on my Spam folder to train Sa, but still they're getting through. There are two separate sets of full headers x-d out to avoid giving server information, but revealing the spam stuff. If anyone has any suggestions on how to tighten this I'd like to know.
> X-Spam-Status: No, score=-0.009 tagged_above=-999 required=1.2 tests=[T_RP_MATCHES_RCVD=-0.01, URIBL_BLOCKED=0.001] autolearn=ham autolearn_force=no >No, score=0.826 tagged_above=-999 required=1.2 tests=[HTML_FONT_LOW_CONTRAST=0.001, HTML_MESSAGE=0.1, MPART_ALT_DIFF=0.724, T_REMOTE_IMAGE=0.01, T_RP_MATCHES_RCVD=-0.01, URIBL_BLOCKED=0.001] autolearn=no autolearn_force=no I don't see BAYES_XX in your tests, look likes bayes isn't being used. Make sure you train at least 200 ham and 200 spam, otherwise bayes does not kick in. That being said, I don't see many rules at all. And a required=1.2 is very low. Did you ever run sa-update and restart amavis? Make sure the new rules are being used. And take a look at the URIBL_BLOCKED, one of your rbl lists is blocked. You are doing too many queries (>100k) or you are using a (public) nameserver which does. You can also do a lint, spamassassin -D --lint and check if there are any errors or not installed dependencies. /MJ
