Title: Message
Some additional information to add today... after updating SpamAssassin to 2.60 yesterday, I am now receiving this message when I try to run the nightly cron job to rebuild the bayes db:
 

/etc/cron.daily/sa-learn-rebuild:

Cannot open bayes databases //.spamassassin/bayes_* R/O: tie failed: Inappropriate ioctl for device Cannot open bayes databases //.spamassassin/bayes_* R/W: tie failed: File exists

What am I doing wrong?  Am I teaching my bayes filters incorrectly......?
 
k
-----Original Message-----
From: Kevin Hanser
Sent: Thursday, November 06, 2003 13:02
To: [EMAIL PROTECTED]
Subject: [SAtalk] questions about bayes filters in spamassassin

Hello, I have spamassassin installed with MailScanner on a redhat linux 9 box, acting as a spam/virus relay.  The mail is not ever stored on the spam relay box, it just scans for viruses and spam, and then forward the mail on to the appropriate mail server (it serves multiple domains).
 
I have a couple questions about the bayes filters in spamassassin, and want to check to make sure I'm doing things right.  I've gone over the docs regarding the bayes filters, and the salearn program, and I currently have set up 2 email addresses at a particular site for people to send misclassified email to:
 
spam:           |/etc/smrsh/spam-learn
ham:            |/etc/smrsh/ham-learn
the two scripts referenced just do a simple call to sa-learn:
spam-learn: sa-learn --spam --no-rebuild --single
ham-learn:  sa-learn --ham --no-rebuild --single
 
And then I have a cron job set to rebuild the bayes db every night:
/usr/bin/sa-learn --rebuild
 
 
I have been instructing users to send misclassifed emails to the spam or ham address using the Outlook "resend" method that was given in the docs.
 
My questions are:
 
1) Is this the correct way for me to have things set up to teach spamassassin about new/misclassified spam?
2) Is there a way to do a similar "resend" method using outlook express, as I have not been able to find a way to do it from OE yet..?
 
 
k

Reply via email to