Can you please check my SA kit to assure me that sa-learn is having the
intended effect? I have ZERO (none, nil) instances where header shows
"autolearn=" as any other value than "autolearn=no". This leads me
conclude that my sa-learn data is not being utilized by spamd?!
Is there a log or other way to peek into SA to have it tell me whether
it's got the sa-learn data in the bayes engine and IS USING THAT
INFORMATION?
*********************************
RedHat AS 4, spamd , qmail, qmail-scanner, clamav
*********************************
spamd -V
SpamAssassin Server version 3.1.7
running on Perl 5.8.5
with SSL support (IO::Socket::SSL 1.02)
************************************
ps waux | grep -i spam
root 3842 0.0 2.9 34520 30328 ? Ss Nov15 0:14
/usr/bin/spamd -x -H /home/spamd -d -s /var/log/spamd/spamd.log
root 26510 0.0 4.0 45476 41316 ? S Dec01 0:38 spamd child
root 31622 0.0 3.2 36920 32688 ? S 11:38 0:00 spamd child
*************************************
sa-learn --dump magic
0.000 0 3 0 non-token data: bayes db version
0.000 0 11737 0 non-token data: nspam
0.000 0 23457 0 non-token data: nham
0.000 0 142588 0 non-token data: ntokens
0.000 0 1164208327 0 non-token data: oldest atime
0.000 0 1165185233 0 non-token data: newest atime
0.000 0 1165184550 0 non-token data: last journal
sync atime
0.000 0 1164928829 0 non-token data: last expiry atime
0.000 0 720391 0 non-token data: last expire
atime delta
0.000 0 29443 0 non-token data: last expire
reduction count
**************************************
ls -l /home/spamd/.spamassassin/
total 7544
-rw------- 1 root root 13248 Dec 3 16:55 bayes_journal
-rw------- 1 root root 5124096 Dec 3 16:33 bayes_seen
-rw------- 1 root root 5398528 Dec 3 16:33 bayes_toks
*************************************
cat /etc/mail/spamassassin/local.cf | grep -v "^#"
required_score 6
rewrite_header Subject [SPAM]
report_safe 0
use_pyzor 0
use_razor2 1
use_dcc 0
dcc_home /var/dcc
skip_rbl_checks 0
rbl_timeout 3
score RCVD_IN_BL_SPAMCOP_NET 2
use_bayes 1
bayes_auto_learn 1
bayes_path /home/spamd/.spamassassin
************************************
I don't see anything here that prevents autolearn results from being
applied?
HELP please! Most appreciated!
Dave.
