Hello all, I sent an email a few days ago but it might not have gotten through 
because I
did not have a PTR entry for my domain and the mailing list requires it.

The problem that I am having is that spambayes is not effective in identifying 
spam on
my new server. On my old server with a totally different setup it worked 
perfectly. Here
is a description of the new setup.

I am using Debian Sarge + postfix + Maildrop with spambayes. I have virtual 
users which
are handeled by the real user vmail. Each user has their own maildroprc file 
(similar to
procmailrc) which uses that specific user's spambayes db file.

I originally copied all the old db files over from my old server but I noticed 
more spam
getting to the unsure and INBOX folders. I noticed that the daily training was 
not
changing the size of the db file at all.

I removed the db file for my user and created a new one with "sb_filter.py -n -d
/home/vmail/.spambayes/dhaval.db". This file turned out to be 12288 bytes. I 
then
proceed to train this db file with my INBOX of  1379 messgaes and spam folder 
with  1336
messages.

I would think that after this many messages of fairly recent emails, I would 
have a very
effective db to filter spam. But the problem is that many spam emails still end 
up in my
unsure and INBOX.

I have read that you can adjust the thresholds, I am just using the defaults 
right now
as this is what I did on the old server and it worked very well.

Can anybody suggest something to help me troubleshoot this problem?


Thanks,
Dhaval
_______________________________________________
[email protected]
http://mail.python.org/mailman/listinfo/spambayes
Check the FAQ before asking: http://spambayes.sf.net/faq.html

Reply via email to