Learning Spam in ISPConfig 3 Question

In step 11 of: The Perfect Server - Debian Squeeze (Debian 6.0) With BIND & Dovecot in ISPConfig 3, the last step reads:

If I follow this howto, does it still work, or do i need to re-enable SpamAssassin?
Learning Spam With SpamAssassin And All Of Your ISPConfig Clients in ISPConfig 3

 

I have been using the Learning Spam setup since this post in April, and it doesn't seem to be learning anything. Even though when I run the shell script it says that is has learned tokens, I keep getting the same spam in my inbox...

Is there something else that I should be doing for it to eliminate the spam emails?

 

I initially read "score" as "core"... Anyway, I don't know what score that I'm using. The email filter is set to "Trigger happy".

 

Till, is this what you needed to answer the question, or is it something else?

 

I found what I think is a pretty good document on this here: http://www200.pair.com/mecham/spam/amavisd-settings.html

 

#!/bin/bash
/usr/bin/sa-learn --spam /var/vmail/*/*/.Junk/cur/*
/usr/bin/sa-learn --ham /var/vmail/*/*/.good/cur/*


Mine is

#!/bin/bash
/usr/bin/sa-learn --spam /var/vmail/biggdog.biz/babedh-dhra/Maildir/.Junk/cur/*
/usr/bin/sa-learn --ham /var/vmail/biggdog.biz/babedh-dhra/Maildir/cur/*

chmod o+x /bin/sa_learn

sa_learn

Thus adding ??
bayes_path /var/lib/amavis/.spamassassin/bayes

If we were to use this script shouldn't we also have to reload amavis?

Now I am not using the cron job simply because I am hand picking my spam and ham and I MUST make sure all of the emails are indeed spam or ham.