Thanks for suggestions Michael, Robert,
But implementing Michael's suggestion is very problematic here -
As you can see below I've got a lot of real spam tagged (hits=) 5.1,5.2,5.3:
--------------------------------------------------------------------
# grep 'X-Spam-Status: Yes, hits=5' spam/*
spam/2140.:X-Spam-Status: Yes, hits=5.0 tag1=-999.0 tag2=5.0 kill=5.0
tests=BAYES_56,
spam/2142.:X-Spam-Status: Yes, hits=5.9 tag1=-999.0 tag2=5.0 kill=5.0
tests=BAYES_60,
spam/2144.:X-Spam-Status: Yes, hits=5.9 tag1=-999.0 tag2=5.0 kill=5.0
tests=BAYES_44,
spam/2197.:X-Spam-Status: Yes, hits=5.8 tag1=-999.0 tag2=5.0 kill=5.0
tests=BAYES_44,
spam/2201.:X-Spam-Status: Yes, hits=5.9 tag1=-999.0 tag2=5.0 kill=5.0
tests=BAYES_30,
spam/2204.:X-Spam-Status: Yes, hits=5.2 tag1=-999.0 tag2=5.0 kill=5.0
tests=BAYES_10,
spam/2208.:X-Spam-Status: Yes, hits=5.8 tag1=-999.0 tag2=5.0 kill=5.0
tests=BAYES_01,
spam/2212.:X-Spam-Status: Yes, hits=5.8 tag1=-999.0 tag2=5.0 kill=5.0
tests=BAYES_60,
spam/2248.:X-Spam-Status: Yes, hits=5.1 tag1=-999.0 tag2=5.0 kill=5.0
tests=BAYES_50,
spam/2279.:X-Spam-Status: Yes, hits=5.1 tag1=-999.0 tag2=5.0 kill=5.0
tests=BAYES_50,
spam/2283.:X-Spam-Status: Yes, hits=5.1 tag1=-999.0 tag2=5.0 kill=5.0
tests=BAYES_50,
spam/2284.:X-Spam-Status: Yes, hits=5.1 tag1=-999.0 tag2=5.0 kill=5.0
tests=BAYES_50,
spam/2286.:X-Spam-Status: Yes, hits=5.1 tag1=-999.0 tag2=5.0 kill=5.0
tests=BAYES_50,
spam/2293.:X-Spam-Status: Yes, hits=5.1 tag1=-999.0 tag2=5.0 kill=5.0
tests=BAYES_50,
spam/2294.:X-Spam-Status: Yes, hits=5.1 tag1=-999.0 tag2=5.0 kill=5.0
tests=BAYES_50,
spam/2295.:X-Spam-Status: Yes, hits=5.1 tag1=-999.0 tag2=5.0 kill=5.0
tests=BAYES_50,
spam/2297.:X-Spam-Status: Yes, hits=5.0 tag1=-999.0 tag2=5.0 kill=5.0
tests=BAYES_44,
spam/2307.:X-Spam-Status: Yes, hits=5.1 tag1=-999.0 tag2=5.0 kill=5.0
tests=BAYES_50,
spam/2313.:X-Spam-Status: Yes, hits=5.3 tag1=-999.0 tag2=5.0 kill=5.0
tests=CLICK_BELOW,
spam/2365.:X-Spam-Status: Yes, hits=5.6 tag1=-999.0 tag2=5.0 kill=5.0
tests=BAYES_90,
And there a lot of ham tagged with 5.1,5.2,...
# grep 'X-Spam-Status: Yes, hits=5' ham/*
ham/1.:X-Spam-Status: Yes, hits=5.7 tag1=-999.0 tag2=5.0 kill=5.0
tests=HTML_50_60,
ham/10.:X-Spam-Status: Yes, hits=5.4 tag1=-999.0 tag2=5.0 kill=5.0
tests=BAYES_44,
ham/17.:X-Spam-Status: Yes, hits=5.6 tag1=-999.0 tag2=5.0 kill=5.0
tests=BAYES_44,
ham/18.:X-Spam-Status: Yes, hits=5.6 tag1=-999.0 tag2=5.0 kill=5.0
tests=BAYES_44,
ham/19.:X-Spam-Status: Yes, hits=5.9 tag1=-999.0 tag2=5.0 kill=5.0
ham/21.:X-Spam-Status: Yes, hits=5.9 tag1=-999.0 tag2=5.0 kill=5.0
tests=BAYES_50,
ham/22.:X-Spam-Status: Yes, hits=5.9 tag1=-999.0 tag2=5.0 kill=5.0
tests=BAYES_44,
ham/25.:X-Spam-Status: Yes, hits=5.1 tag1=-999.0 tag2=5.0 kill=5.0
ham/27.:X-Spam-Status: Yes, hits=5.6 tag1=-999.0 tag2=5.0 kill=5.0
tests=BAYES_44,
ham/28.:X-Spam-Status: Yes, hits=5.7 tag1=-999.0 tag2=5.0 kill=5.0
tests=BAYES_99,
ham/30.:X-Spam-Status: Yes, hits=5.2 tag1=-999.0 tag2=5.0 kill=5.0
ham/31.:X-Spam-Status: Yes, hits=5.2 tag1=-999.0 tag2=5.0 kill=5.0
tests=BAYES_01,
ham/32.:X-Spam-Status: Yes, hits=5.2 tag1=-999.0 tag2=5.0 kill=5.0
tests=BAYES_30,
ham/33.:X-Spam-Status: Yes, hits=5.5 tag1=-999.0 tag2=5.0 kill=5.0
tests=BAYES_00,
ham/35.:X-Spam-Status: Yes, hits=5.1 tag1=-999.0 tag2=5.0 kill=5.0
tests=BAYES_44,
ham/38.:X-Spam-Status: Yes, hits=5.9 tag1=-999.0 tag2=5.0 kill=5.0
tests=BAYES_30,
ham/39.:X-Spam-Status: Yes, hits=5.4 tag1=-999.0 tag2=5.0 kill=5.0
tests=BAYES_99
ham/4.:X-Spam-Status: Yes, hits=5.6 tag1=-999.0 tag2=5.0 kill=5.0
tests=HTML_MESSAGE,
ham/40.:X-Spam-Status: Yes, hits=5.5 tag1=-999.0 tag2=5.0 kill=5.0
tests=BAYES_00,
ham/41.:X-Spam-Status: Yes, hits=5.5 tag1=-999.0 tag2=5.0 kill=5.0
tests=BAYES_00,
ham/42.:X-Spam-Status: Yes, hits=5.5 tag1=-999.0 tag2=5.0 kill=5.0
tests=BAYES_00,
ham/43.:X-Spam-Status: Yes, hits=5.5 tag1=-999.0 tag2=5.0 kill=5.0
tests=BAYES_00,
ham/44.:X-Spam-Status: Yes, hits=5.1 tag1=-999.0 tag2=5.0 kill=5.0
tests=BAYES_50,
ham/7.:X-Spam-Status: Yes, hits=5.7 tag1=-999.0 tag2=5.0 kill=5.0
tests=BAYES_50,
------------------
Even if I tech BAYESIAN DB as amavis user and try to resend "ham" message, SA
keeps identifying it as spam and put it in quarantine.
For example:
------------
# sendmail -itf [EMAIL PROTECTED] < /var/spool/imap/user/spamcop/ham/17.
# grep [EMAIL PROTECTED] /var/log/amavis.log Nov 7 11:45:20
mail.mydomain.ac.il amavisd[10839]: (10839-06) ESMTP::10024
/var/spool/amavis/amavis-20061107T114214-10839: <[EMAIL PROTECTED]> -> <[EMAIL
PROTECTED]> Received: SIZE=5467 from mail.mydomain.ac.il ([127.0.0.1]) by
localhost (mail.mydomain.ac.il [127.0.0.1]) (amavisd-new, port 10024) with
ESMTP id 10839-06 for <[EMAIL PROTECTED]>; Tue, 7 Nov 2006 11:45:20 +0200
(IST) Nov 7 11:45:20 mail.mydomain.ac.il amavisd[10839]: (10839-06) Checking:
<[EMAIL PROTECTED]> -> <[EMAIL PROTECTED]> Nov 7 11:45:23 mail.mydomain.ac.il
amavisd[10839]: (10839-06) SPAM, <[EMAIL PROTECTED]> -> <[EMAIL PROTECTED]>,
Yes, hits=5.6 tag1=-999.0 tag2=5.0 kill=5.0 tests=BAYES_50,
FORGED_HOTMAIL_RCVD2, HTML_FONT_BIG, HTML_MESSAGE, MIME_HEADER_CTYPE_ONLY,
MIME_HTML_ONLY, NO_REAL_NAME, quarantine
spam-47d276560a19b33a3ea18035405e4804-20061107-114523-10839-06 ([EMAIL
PROTECTED])
Nov 7 11:45:23 mail.mydomain.ac.il amavisd[10839]: (10839-06) BAD HEADER from
<[EMAIL PROTECTED]>: Improper use of control character (char 0D hex) in message
header 'Received'\n Received: ...haifa.ac.il ([unix socket])\\r\\n\\tby
mail.edu.h...\n ^
Nov 7 11:45:23 mail.mydomain.ac.il amavisd[10839]: (10839-06) Not-Delivered,
<[EMAIL PROTECTED]> -> <[EMAIL PROTECTED]>, quarantine
spam-47d276560a19b33a3ea18035405e4804-20061107-114523-10839-06, Message-ID:
<[EMAIL PROTECTED]>, Hits: 5.577
These are the headers of this specific mail (Some of these headers, the upper
ones, are from mail relay that is managed by another admin and listed before
mine as MX record in DNS):
--------------------------------------------
Return-Path: <[EMAIL PROTECTED]>
Received: from mail.mydomain.ac.il ([unix socket])
by mail.mydomain.ac.il (Cyrus v2.2.3) with LMTP; Mon, 30 Oct 2006
13:36:55 +0200
X-Sieve: CMU Sieve 2.2
Received: from localhost (localhost [127.0.0.1])
by mail.mydomain.ac.il (Postfix) with ESMTP id 8C9E31C379
for <[EMAIL PROTECTED]>; Mon, 30 Oct 2006 13:36:55 +0200 (IST)
X-Envelope-To: <[EMAIL PROTECTED]>
X-Envelope-From: <[EMAIL PROTECTED]>
X-Quarantine-id:
<spam-88a5a789750e792aafbdb35c638f04fc-20061030-133655-02873-03>
Received: from mr3.haifa.ac.il (mr3.haifa.ac.il [132.74.1.219])
by mail.mydomain.ac.il (Postfix) with ESMTP id 51E291F7FE
for <[EMAIL PROTECTED]>; Mon, 30 Oct 2006 13:36:54 +0200 (IST)
Received: from localhost (localhost [127.0.0.1])
by mr3.haifa.ac.il (Postfix) with ESMTP id 20A5019105
for <[EMAIL PROTECTED]>; Mon, 30 Oct 2006 13:20:40 +0200 (IST)
X-Virus-Scanned: amavisd-new at haifa.ac.il
Received: from mr3.haifa.ac.il ([127.0.0.1])
by localhost (mr3.haifa.ac.il [127.0.0.1]) (amavisd-new, port 10026)
with ESMTP id TTzCk2m8o8M9 for <[EMAIL PROTECTED]>;
Mon, 30 Oct 2006 13:20:39 +0200 (IST)
Received: from www.mydomain.ac.il (www.mydomain.ac.il [132.74.41.30])
by mr3.haifa.ac.il (Postfix) with ESMTP id E94E8190F5
for <[EMAIL PROTECTED]>; Mon, 30 Oct 2006 13:20:39 +0200 (IST)
Received: by www.mydomain.ac.il (Postfix, from userid 30)
id 6A04B200E0; Mon, 30 Oct 2006 13:33:24 +0200 (IST)
To: [EMAIL PROTECTED]
Subject: mailing list
From: [EMAIL PROTECTED]
Content-type: text/html; charset=windows-1255
Message-Id: <[EMAIL PROTECTED]>
Date: Mon, 30 Oct 2006 13:33:24 +0200 (IST)
X-Spam-Status: Yes, hits=5.6 tag1=-999.0 tag2=5.0 kill=5.0 tests=BAYES_44,
FORGED_HOTMAIL_RCVD2, HTML_FONT_BIG, HTML_MESSAGE, MIME_HEADER_CTYPE_ONLY,
MIME_HTML_ONLY, NO_REAL_NAME
X-Spam-Level: *****
-----------------------------------
Any Suggestions?
Best Regards,
Leon Kolchinsky
-----Original Message-----
From: [EMAIL PROTECTED] [mailto:[EMAIL PROTECTED] On Behalf Of Muenz, Michael
Sent: Monday, November 06, 2006 5:30 PM
To: [email protected]
Subject: Re: [Web-cyradm] How to resend messages?
> Actually I didn't set the score of BAYES_99. It's the default
> one as I have not changed it, I don't really know its value
> (I believe you say that it is 5.0).
Your kill-level is 5.0 and the score for BAYES_99 is 5.4.
I would twist it, kill-level 5.4 and BAYES_99 to 5.0 :)
_______________________________________________
This mailing list is hosted and supported
by bit-heads GmbH | http://www.bit-heads.ch
_______________________________________________
Web-cyradm mailing list
[email protected]
http://www.web-cyradm.org/mailman/listinfo/web-cyradm
_______________________________________________
This mailing list is hosted and supported
by bit-heads GmbH | http://www.bit-heads.ch
_______________________________________________
Web-cyradm mailing list
[email protected]
http://www.web-cyradm.org/mailman/listinfo/web-cyradm
