BAYES issues

Anthony Peacock a.peacock at chime.ucl.ac.uk
Thu Feb 15 20:38:34 CET 2007


Jay Chandler wrote:
> Anthony Peacock wrote:
>> This message got a BAYES_20 score.  Your SA is using Bayes correctly.
>>
>> Why don't you think it is working?
> 
> aconcagua# cat /var/log/maillog |grep BAYES
> Feb 15 01:24:24 aconcagua MailScanner[57909]: Message A60174554E.EC45D 
> from 65.54.246.232 (bf1997 at hotmail.fr) to chapman.edu is spam, 
> SpamAssassin (not cached, score=8.308, required 6, SARE_BAYES_5x7 0.60, 
> SARE_FRAUD_X3 1.67, SARE_FRAUD_X4 1.67, SARE_FRAUD_X5 1.67, 
> SARE_FRAUD_X6 1.67, SARE_MILLIONSOF 0.32, SARE_URGBIZ 0.72)
> Feb 15 05:44:29 aconcagua MailScanner[98220]: Message 55C7A455C4.48148 
> from 146.142.40.232 (bounce-ximpim-351275 at list.bls.gov) to chapman.edu 
> is not spam, SpamAssassin (not cached, score=3.5, required 6, 
> MANGLED_MEDS 2.30, SARE_BAYES_5x7 0.60, SARE_BAYES_6x7 0.60)
> aconcagua#
> 
> This is for approximately ten thousand messages since six hours ago.
> 
> Do I need a Bayes rule that defines scores for each level?  I'm not sure 
> I have one by default...

In which case I suspect that you have a working Bayes for whichever user 
you run the command line tests as, but not for the user which 
MailScanner runs as.

Which user did you log in as to run the SA command line tests?

Which user does MailScanner run as?

By default SA will use different Bayes databases for different users. 
So if MailScanner runs as root it will use root's Bayes database, if you 
then run the SA command line as user jaychandler it will use that user's 
database.

-- 
Anthony Peacock
CHIME, Royal Free & University College Medical School
WWW:    http://www.chime.ucl.ac.uk/~rmhiajp/
"If you have an apple and I have  an apple and we  exchange apples
then you and I will still each have  one apple. But  if you have an
idea and I have an idea and we exchange these ideas, then each of us
will have two ideas." -- George Bernard Shaw


More information about the MailScanner mailing list