spamassassin not lethal anymore

[RW]

On Thu, 12 Jan 2017 14:58:10 -0700
Russell L. Carter wrote:

> On 01/11/17 16:18, Valeri Galtsev wrote:
> >
> > On Wed, January 11, 2017 4:00 pm, Tim Daneliuk wrote:  
> >> Milter greylist also helps considerably
> >>  
> >
> > I'm usually using sqlgrey for greylisting. In addition to that I'm
> > using amavisd (that harnesses spamassassin, clamav (and/or other
> > anti-viruses, - yes, I have windows people receiving their mail on
> > our UNIX servers), and variety of distributed spam signature
> > databases. Also, in postfix configuration (that we use as MTA) we
> > add some RBLs:
> >
> >         reject_rbl_client zen.spamhaus.org,
> >         reject_rbl_client bl.spamcop.net,
> >         reject_rbl_client cbl.abuseat.org,
 

cbl is part of zen


> > I hope, this helps.  
> 
> I had hoped I wouldn't need to do a deep dive on state-of-the-art
> spam killing, but it's 2017 and the game has gotten tougher.

You haven't actually determined whether you have a specific problem
with SA, or whether your spam has just got harder. You might want to
upload a few low-scoring spams to pastebin and ask on the SA user
list.

> Thanks for these suggestions.  I'm implementing greylisting, pyzor,
> razor,

razor2 has been a standard part of SA and on by default, since 2006. 

>  and the above client_recipient_restrictions.  dcc I don't
> quite understand yet.

be aware that dcc is not a test for spam, it's a test for any kind
of bulk mail.