bsd wrote:

> Hello,

> I have setup amavisd-new to be used with spamassassin & dcc & razor2 //

> In the directory /var/db/.spamassassin I have couple of files  
> including auto_whitelist :

>> root:newmail 8:09 /var/amavis/.spamassassin # ls -alh
>> total 12504
>> drwx------   2 vscan  vscan   512B May  2 08:09 .
>> drwxr-x---  12 vscan  vscan   1.0K May  2 07:00 ..
>> -rw-------   1 vscan  vscan    10M May  2 08:09 auto-whitelist
>> -rw-------   1 vscan  vscan     6B Jan 19  2005 auto-whitelist.mutex
>> -rw-------   1 vscan  vscan   1.0K Jan 19  2005 bayes.mutex
>> -rw-------   1 vscan  vscan    77K Feb 18  2005 bayes_journal
>> -rw-------   1 vscan  vscan   2.5M Feb 18  2005 bayes_seen
>> -rw-------   1 vscan  vscan   2.0M Feb 18  2005 bayes_toks
>> -rw-r--r--   1 vscan  vscan   1.5K Nov 14 17:35 user_prefs


> Though autowhitelist seems to be deactivated both in amavisd.conf &  
> local.cf for spamassassin, the size of these files keeps on growing  
> fast.

> I would like to understand where to tune this parameter and what is
> the goal of this file (and the other bayes...).

The setting in amavisd.conf has no effect when used with newer
versions of spamassassin. If using version 3.1.x of spamassassin,
comment out this line in v310.pre to disable auto-whitelist if desired:
loadplugin Mail::SpamAssassin::Plugin::AWL
as opposed to disabling it in local.cf. Auto-whitelist will grow
forever and I have seen some larger than 700MB. You can use the
check_whitelist tool (in the tools directory of the source code) to
trim down the file somewhat.
Or:  http://www.deepnet.cx/~kdeugau/spamtools/trim_whitelist

See the spamassassin wiki for more info:
http://wiki.apache.org/spamassassin/AutoWhitelist

The bayes files will grow to somewhere around 7MB and will clean
themselves up automatically. The bayes files are a database of tokens
(pieces of information contained in email messages) that are used to
decide if a given message is more "spam like" or "not spam like". It
is set to automatically learn which is which by being fed spam and
ham (not spam) but the 'sa-learn' program can be used to feed it
manually. Its decision affects the total spam score:

X-Spam-Status: No, score=-0.587-9 tagged_above=-999 required=5
        tests=[BAYES_00=-2.599, BIZ_TLD=2.013, SPF_PASS=-0.001]
               ^^^^^^^^

Read the documentation:
               
http://wiki.apache.org/spamassassin/BayesFaq
http://spamassassin.apache.org/full/3.1.x/dist/doc/
http://spamassassin.apache.org/full/3.1.x/dist/doc/sa-learn.html

Gregober --->> PGP ID --> 0x1BA3C2FD

Gary V



-------------------------------------------------------
Using Tomcat but need to do more? Need to support web services, security?
Get stuff done quickly with pre-integrated technology to make your job easier
Download IBM WebSphere Application Server v.1.0.1 based on Apache Geronimo
http://sel.as-us.falkag.net/sel?cmd=lnk&kid=120709&bid=263057&dat=121642
_______________________________________________
AMaViS-user mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/amavis-user
AMaViS-FAQ:http://www.amavis.org/amavis-faq.php3
AMaViS-HowTos:http://www.amavis.org/howto/

Reply via email to