Print Page | Close Window

use corpus.db from more than one SF

Printed From: LogSat Software
Category: Spam Filter ISP
Forum Name: Spam Filter ISP Support
Forum Description: General support for Spam Filter ISP
URL: https://www.logsat.com/spamfilter/forums/forum_posts.asp?TID=6126
Printed Date: 28 December 2024 at 7:49pm


Topic: use corpus.db from more than one SF
Posted By: kp4711
Subject: use corpus.db from more than one SF
Date Posted: 28 June 2007 at 2:54pm

It is possible to use the corpus.db Files from more than one SF-installation to share the informations in the files?

 


Replies:
Posted By: LogSat
Date Posted: 28 June 2007 at 4:13pm
You may be able to copy the files from one installation to the other, but we recommend against doing that, unless the servers are load-balanced and receive the same amount and quality of traffic (You will need to stop SpamFilter on the server where you are copying the file to before copying the file).

This is because the bayesian database works and learns spam by examining the incoming emails. If the email traffic actually received by SpamFilter does not reflect the email traffic used to create the database, you will get very inaccurate results. For example, if you have two servers on a primary and secondary MX records, the server on the primary will receive a mix of spam and good emails. The one on your secondary will receive mostly spam, as spammers often send spam emails directly to all MX records. Thus the type of traffic they receive is very much different, and the bayesian statistical data in the database on each server will also be very different. Mixing the two will yield unpredictable (i.e. inaccurate) results.


-------------
Roberto Franceschetti

http://www.logsat.com" rel="nofollow - LogSat Software

http://www.logsat.com/sfi-spam-filter.asp" rel="nofollow - Spam Filter ISP


Print Page | Close Window