spamassassin-users mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Sergio Durigan Junior <sergi...@sergiodj.net>
Subject SA not correctly classifying spam
Date Mon, 11 Nov 2013 22:26:48 GMT
Hi there,

As requested by Karsten here (I took the liberty to include him in the
Cc list):

   <https://mail-archives.apache.org/mod_mbox/spamassassin-users/201311.mbox/browser>

I am starting this new thread in order to try to solve/identify what's
going on with my SA instance (*if* there's anything wrong, of course).

First of all, I am using:

- Debian 7.1 (stable)

- SpamAssassin version 3.3.2
    running on Perl version 5.14.2

Here is an example of a misclassified spam message:

   <http://sergiodj.net/~sergio/sa/spam.txt>

(This spam message was sent to a mailing list, not directly to my
address, as can be seen.  I still don't have spams that were sent
directly to my e-mail address.).

And here's a ham:

   <http://sergiodj.net/~sergio/sa/ham.txt>

Here's the content of /etc/spamassassin/local.cf:

   <http://sergiodj.net/~sergio/sa/local.cf.txt>

(As I mentioned in another message, this is Debian's default file,
untouched.).

Here's what I see when I run "sa-learn --dump magic":

   0.000          0          3          0  non-token data: bayes db version
   0.000          0          5          0  non-token data: nspam
   0.000          0         71          0  non-token data: nham
   0.000          0       6229          0  non-token data: ntokens
   0.000          0 1383057593          0  non-token data: oldest atime
   0.000          0 1384207955          0  non-token data: newest atime
   0.000          0 1384058847          0  non-token data: last journal sync atime
   0.000          0          0          0  non-token data: last expiry atime
   0.000          0          0          0  non-token data: last expire atime delta
   0.000          0          0          0  non-token data: last expire reduction count

AFAIU "nspam" is much smaller than "nham" because autolearn is enabled.

And here's how I run my spam "solution" on my server:

- I run spamd as root, using the following options:

     --create-prefs --max-children 5 --helper-home-dir
     --allow-tell -d --pidfile=/var/run/spamd.pid

- I run spamc directly from my .procmailrc:

     :0fw: spamassassin.lock
     * < 256000
     | spamc

     # All mail tagged as spam (eg. with a score higher than the set
     # threshold) is moved to "Spam/".
     :0
     * ^X-Spam-Status: Yes
     Spam/

- I update SA rules (sa-update) daily via cronjob.

- I feed every spam message that I receive to "sa-learn" (however, now I
  am keeping the spam messages around in order to diagnose the problem).

I am wondering what could possibly be wrong in my configuration.  Maybe
I should tweak the SA's config files more, in order to get a proper
detection done.  Suggestions and comments are welcome, of course.

Thanks,

-- 
Sergio

Mime
View raw message