Hi,
I’m a newbie. I built a Postfix –Spamassassin-Cyrus
IMAP mail server using Luc de Louw’s How To at http://www.delouw.ch/linux/postfix.phtml.
Recently I added SpamAssassin and Anomy Sanitizer to the system by following: http://advosys.ca/papers/postfix-filtering.html.
This too was a success.
I notice that spam continues to get by SpamAssasin so I would like to
train the Bayesian Classifier by forwarding both SPAM and HAM to special
mailboxes configured in Cyrus for this purpose. I added two users (spam
and ham) to Cyrus and forwarded mail to the boxes. I then
attempted running sa-learn on the mail boxes but could not get it to read them.
The mailboxes I attempted to read are in /var/spool/imap/user/us003 and
/var/spool/imap/user/us004 (the accounts configured for spam and ham respectively).
I tried with all the possible sa-learn switches for mailbox types (--dir,
--file –mbox) but none worked. Am I doing this correctly?
Please help.
Thanks,
Jeff
|