You are viewing a plain text version of this content. The canonical link for it is here.
Posted to users@spamassassin.apache.org by Alexander Gruber <sy...@web.de> on 2012/06/17 14:24:21 UTC

bayes filter: all mails are BAYES_00

Hello,

I experience some strange problem with the bayesian filter from
spamassassin. Absolutly all mails are tagged with BAYES_00 and therefore
the score is lowered. Especially with spam mails this is anoying. Any
ideas why this is happening? I have disabled the per-user bayes in favor
of one big site-wide database by the way.

Here is my bayes relevant configuration. autolearn is the default "yes".
Learning thresholds are the default ones.

bayes_store_module              Mail::SpamAssassin::BayesStore::SQL
bayes_sql_dsn                       DBI:mysql:spamassassin:localhost
bayes_sql_override_username     vpopmail
bayes_expiry_max_db_size        2000000


sa-learn --dump magic
0.000          0          3          0  non-token data: bayes db version
0.000          0    1405419          0  non-token data: nspam
0.000          0     511568          0  non-token data: nham
0.000          0     221395          0  non-token data: ntokens
0.000          0 1338655736          0  non-token data: oldest atime
0.000          0 1339935261          0  non-token data: newest atime
0.000          0          0          0  non-token data: last journal
sync atime
0.000          0 1339708206          0  non-token data: last expiry atime
0.000          0     345600          0  non-token data: last expire
atime delta
0.000          0       7500          0  non-token data: last expire
reduction count

Re: bayes filter: all mails are BAYES_00

Posted by RW <rw...@googlemail.com>.
On Sun, 17 Jun 2012 14:24:21 +0200
Alexander Gruber wrote:

> Hello,
> 
> I experience some strange problem with the bayesian filter from
> spamassassin. Absolutly all mails are tagged with BAYES_00 and
> therefore the score is lowered. 

> bayes_expiry_max_db_size        2000000
 
> 0.000          0     221395          0  non-token data: ntokens

> 0.000          0 1339708206          0  non-token data: last expiry

One thing that stands-out is that you have  2,000,000 tokens
configured, but a few days ago an expiry left you with fewer than
221,395. Unless this is a very recent configuration change it looks
like your configuration isn't being picked-up.