I setup my current mail servers +/- 2 years ago. Debian wheezy, amavis 
(calls spamassassin and dspam), postfix. Postgres 9.1 database. After 
the initial setup I train dspam with a few thousand email (ham and 
spam). It all works well.

I'm using dspam in this kind of configurations for many years. I 
normally create a cronjob which purges the database (I use purge.sql, 
part of the source). This time I forgot the cronjob. A few weeks ago I 
noticed the huge size of the dspam database and decided to purge the 
database.

The database was enormous reduced. Afterwards spam was detected as 
before but quite a annoying lot of false positives. Normal clear 
innocent mail was rejected by dspam. Amavis log files show in such a 
case something like this:
DSPAM result: Spam, score=10.000, sig=
DSPAM: suppl attr: X-DSPAM-Result = 'Spam'
DSPAM: suppl attr: X-DSPAM-Class = 'Spam'
DSPAM: suppl attr: X-DSPAM-Confidence = '0.50'
DSPAM: suppl attr: X-DSPAM-Probability = '0.9758'
DSPAM: suppl attr: X-DSPAM-Signature = 'N/A'

My solution is to retrain dspam with a lot of ham. That seems to help 
but I do not understand why dspam is so wrong after a (very late) purge.

Anyone a suggestion?

-- 
Regards,

Ruud Baart


------------------------------------------------------------------------------
"Accelerate Dev Cycles with Automated Cross-Browser Testing - For FREE
Instantly run your Selenium tests across 300+ browser/OS combos.
Get unparalleled scalability from the best Selenium testing platform available
Simple to use. Nothing to install. Get started now for free."
http://p.sf.net/sfu/SauceLabs
_______________________________________________
Dspam-user mailing list
Dspam-user@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/dspam-user

Reply via email to