Bowie Bailey <Bowie_Bailey <at> BUC.com> writes: > > > all rules seem the same, expected BAYES_50 vs BAYES_99. What will > > the causing for this value so high at backup server. Also, i had set > > internal_networks 192.168.2.0/24 at both spamassassin local.cf at > > both server, will it reduce/increase the BAYES_ value? > > The Bayes rules depend on exactly two things: > > 1) The training of the Bayes database > 2) The content of the email > > No other settings will affect them. > > If you have a ham message that is getting a BAYES_99 rule hit, then > the Bayes database for that user needs to be retrained. >
the spamassasin had disabled personal bayes database and both mail server strained similar spam mail. So i wonder why the backup one will socred the mail as BAYES_99 (mean that the bayes database think that this mail is 99% spam mail?)
