cnone wrote: > I have some mails that I know they are spam but spamassassin gives secores > below 5.0(generally zero) for some of them. I updated the rules,changed the > score threshold but spamassassin still sees them as normal emails. Am I > missing something or is this normal? I have like 1800 emails but it sees > only 5 of them as spam. > Well, you could start off by training the bayes database to know that they are spam, making use of the sa-learn tool to do so.
$man sa-learn You might also want to make sure none of those messages are matching ALL_TRUSTED. If any do, then you probably need to configure your trusted_networks manually. (The trust-path auto-guesser gets confused if your MX is NATed, or otherwise has a non-routable IP)
