Betr.: Bayes problem

[Arjan Melein]

>>> Op 31-5-2012 om 11:49 is door Nenad Vijatov <nenad.vijatov at gmail.com>
geschreven:
> Hi guys,
> 
> I have a problem with spamassassin bayes autolearn, exactly the problem is
> I have different bayed db:
> 
> sa-learn --dump magic:
> 0.000          0          3          0  non-token data: bayes db version
> 0.000          0         27          0  non-token data: nspam
> 0.000          0          0          0  non-token data: nham
> 0.000          0      28179          0  non-token data: ntokens
> 
> sa-learn --dump magic -p /etc/MailScanner/spam.assassin.prefs.conf:
> 0.000          0          3          0  non-token data: bayes db version
> 0.000          0         89          0  non-token data: nspam
> 0.000          0        135          0  non-token data: nham
> 0.000          0      21025          0  non-token data: ntokens
> 
> Is this a problem with wrong configuration, does my spamassassin use
> spam.assassin.prefs.conf or not? And how I can be sure that my spamassassin
> uses spam.assassin.prefs.conf or not, and which one of those two dumps are
> the right one that spamassassin use?
> 
> 
> Thank you & Cheers!

Usually having different bayes db's is caused by spamassassin running as a different user then that you're running the command as.
Spamassassin will probably run as the same user that MS is running at, if you have your bayes in MySQL this is easy to check by looking at the tables.

I have overridden the bayes MySQL user by putting 'bayes_sql_override_username       postfix' in /etc/mail/spamassassin/init.pre

-
Arjan