How to optimize Spamassassin on shared server?

Hi folks,

I have been experimenting with Spamassassin settings on my Nocster shared servers lately, trying to optimize how much spam it catches.

I have read up on how Spamassassin can access different realtime blacklists such as DCC and Pyzor, but can't seem to make it use them.

I have added the following to my user_prefs file in .Spamassassin:



Code: 
use_dcc 1  
use_pyzor 1  
use_razor2 1  
use_bayes 1  
auto_learn 1  
skip_rbl_checks 0

That should be turning ON all those features, but it doesn't seem to be. You can see what Spamassassin is doing by adding a custom header to user_prefs:


Code: 
  add_header all Status _YESNO_, hits=_HITS_ required=_REQD_ tests=_TESTS_ bayes=_BAYES_ awl=_AWL_ autolearn=_AUTOLEARN_ version=_VERSION_ dcc results=_DCCR_  pyzor results= _PYZOR_  rblresults=_RBL_

The header that line generates looks like this (for example):
X-Spam-Status: Yes, hits=9.8 required=4.2 tests=DATE_SPAMWARE_Y2K,HTML_40_50, HTML_MESSAGE,HTML_MIME_NO_HTML_TAG,MIME_HTML_NO_CH
ARSET, MIME_HTML_ONLY,MIME_HTML_ONLY_MULTI,MISSING_MIMEOL
E, MISSING_OUTLOOK_NAME bayes=0.5 awl= autolearn=no version=2.60 dcc results= pyzor results= rblresults=

You can see that dcc, pyzor, and rbl have no results.

Is blacklist access turned off for the Nocster boxes because it is more resource intensive?

Any way to turn it ON on a site by site basis?

Anyone have other Spamassassin tips?

Anyone know how to use Bayes or Autolearn? I see that Bayes is building some sort of database on my server, but it is not being used.

 

 

 

 

Top