Originally Posted by falko
Thank you Falko, my server does catch spam but is letting through an awfull lot that has words I am surprised it lets through.
I have played with the 'spam score' which helped a bit.
How do I check that auto learning is on and working properly?
I have tried following the details in the link you provide.
I ran; 'sa-learn --spam /var/www/web1/user/web1_user/Maildir/.Junk
as well as the 'sa-learn --ham' command but with the same result;
'bash: sa-learn: command not found'
I have read that the sa-learn is part of the SpamAssassin perl module, so I assume I have it but the system thinks not!
Where am I going wrong, do I need to be in a specific directory when I run the command?
Also you say that it should self learn without intervention, how can it do this, does it get info back from Roundcubemail, if so how?