Spamassassin, Bayesian Filtering

[Hamish]

On Wednesday 16 November 2005 14:23, Markus F. Kraemer wrote:
> Hi List,
>
> I am playing around with Kolab 2.0.1 on a Suse 9.3 system since a while.
> Everything is working like a charm except Spamassassin´s Bayesian
> Filtering. I followed the instructions in the wiki
> (http://wiki.kolab.org/index.php/Fighting_spam) and trained my DB with
> ham and spam. But the bayesian filtering doesn´t seem to work.
>
> Is it supposed to work out of the box or do I have to make some changes
> in the config files?
>
> Which user do I have to use for training?
> Even though I use the --dbpath <path> option for the "sa-learn" command
> and the result value of sa-learn tells me that he successfully learned
> from n messages the nspam value of a "sa-learn -D --dump magic" is not
> changing.
>
> Can someone give me a hint where to continue debugging?? Let me know if
> you need any additional information!
>
> Cheers,
>
> Markus

Hi Markus
I apologise if this seems very simple, but are you definately using kolab's 
sa-learn? (`which sa-learn`) I made the mistake of using the hosts version 
and also wondered what was happening. To make things easier, try the opa 
script http://wiki.kolab.org/index.php/Opa . 
What are you using for <dbpath>?
I have spamassassin working very nicely after some training, I did change the 
amavisd template a bit though (lowered the minimum score for marking and 
killing)
Hope that is some help,
Hamish
-------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Type: application/pgp-signature
Size: 189 bytes
Desc: not available
URL: <http://lists.kolab.org/pipermail/users/attachments/20051116/d6fe9512/attachment.sig>