[SAtalk] Auto Learn Question

[Eyvonne Sharp]

I am using SA 2.55 on a mail server for only a few select users.  We have auto_learn turned on for those users using procmail (I am not running the daemon).       .procmailrc ########################################################################### :0fw | /usr/bin/spamassassin -P -a :0 * ^X-Spam-Status: Yes $HOME/Spam ##################################   I haven't been able to find any information in the docs on what messages are auto_learned and why.  Users that have received over thousands of emails and have had hundreds of messages classified as spam will only have a bayes_msgcount of 94.  Does that mean that auto_learn has only 'read' 94 email messages?  If this is the case, it isn't very helpful.   For users that we manually execute sa-learn for, their msgcount is close to the number of messages run through sa-learn and the Bayesian filtering works well.   How do I get auto_learn to recognize more messages or even all messages that are run through SA?   Please advise.   user_prefs ########################################################################### # How many hits before a message is considered spam. required_hits           5.0   # Whether to change the subject of suspected spam rewrite_subject         0   # Text to prepend to subject if rewrite_subject is used subject_tag             *****SPAM*****   # Encapsulate spam in an attachment report_safe             1   # Use terse version of the spam report use_terse_report        0   # Enable the Bayes system use_bayes               1   # Enable Bayes auto-learning auto_learn              1   # Enable or disable network checks skip_rbl_checks         0 use_razor2              1 use_dcc                 1 use_pyzor               1   # Mail using languages used in these country codes will not be marked # as being possibly spam in a foreign language. ok_languages            all   # Mail using locales used in these country codes will not be marked # as being possibly spam in a foreign language. ok_locales              all   ##################################       Eyvonne Sharp Web Services Manager Box Lake Networks (859) 737-4400 http://www.boxlake.com