I have had great success with a home-grown anti-spam system using
- sendmail
- some RBLs
- scripts to search our sendmail logs for addresses we have sent to, since
they are presumably people we want to hear from
- scripts to search our LDAP accessible (Notes) database of customers, who
we also love to hear from.
All this data is in sendmail's /etc/mail/access.db.
These whitelisting processes were the key to eliminating false positives
being reported by users. The key is that if you send someone mail, you
probably want to hear back. Plus users often solve their own problems - if
someone sends us mail that gets bounced, then our user will often try to
send them mail as a test, which of course whitelists the original sender,
and eliminates the call to our help desk.
I have added MimeDefang and SpamAssassin, and am getting even more spam
caught. Thank goodness, since we just lost Osirusoft!
Is there any way to get a SpamAssassin rule to check /etc/mail/access.db?
For now, I am generating a /etc/mail/spamassassin/whitelist.cf, but that is
a 20,000 line file. With 600 users, this seem like a reasonable number
reasonable to me.
(OT: Does anyone know if this whitelist.cf is being reread each message, or
only each time I run "service mimedefang reread".)
Wrolf Courtney
Donovan Data Systems, Inc.
(212) 633-5470
-------------------------------------------------------
This sf.net email is sponsored by:ThinkGeek
Welcome to geek heaven.
http://thinkgeek.com/sf
_______________________________________________
Spamassassin-talk mailing list
[EMAIL PROTECTED]
https://lists.sourceforge.net/lists/listinfo/spamassassin-talk
