AHA! That brings me to my other question on this mailing list "Which DB is
actually used?".
General rule for learning: learn as the user you use to filter mail.
Corolary: don't learn as root, since SA never runs as root.
You can generally set up SA two different ways: site-wide bayes, or
individual user bayes. How you do either probably depends in part on what
the rest of your mail system is.
It isn't quite clear to me how you have your system set up. In any case I
don't know enough to tell you exactly what you need to do to get either
effect; I could only suggest RTFM, which doesn't always help much.
It looks to me like the exim user is probably the user that you main mail
processing is running under. I'd say that was the main database and you
were set up site-wide, but that doesn't explain why there is a second
database under your usercode.
Perhaps the exim user is the main filter and has been doing auto-learning,
and you have also been manually learning mails under your own usercode? But
the database sizes are relatively even, so maybe you somehow are processing
mail under two different usercodes for different recipients?
Loren
I did sa-learn --dump magic as root, but root has never been tought any
spam or ham.
I did "sa-learn --dump magic --dbpath ...":
max:~# sa-learn --dump magic --dbpath /home/bosse/.spamassassin/
0.000 0 3 0 non-token data: bayes db version
0.000 0 2447 0 non-token data: nspam
0.000 0 1320 0 non-token data: nham
0.000 0 152520 0 non-token data: ntokens
0.000 0 1145221056 0 non-token data: oldest atime
0.000 0 1157497564 0 non-token data: newest atime
0.000 0 0 0 non-token data: last journal sync
atime
0.000 0 1157464926 0 non-token data: last expiry atime
0.000 0 11059200 0 non-token data: last expire atime
delta
0.000 0 21382 0 non-token data: last expire
reduction count
max:~# sa-learn --dump magic --dbpath /var/spool/exim4/.spamassassin/
0.000 0 3 0 non-token data: bayes db version
0.000 0 477 0 non-token data: nspam
0.000 0 1966 0 non-token data: nham
0.000 0 69851 0 non-token data: ntokens
0.000 0 1130849621 0 non-token data: oldest atime
0.000 0 1157701646 0 non-token data: newest atime
0.000 0 1157687964 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
Since the filtering is working quite well, I guess one of these two
databases are used. The user "bosse" is my own user for emails. The
database in /var/spool/exim4 is that for auto learning?
Questions:
1. How can I know which of these two are actually used for filtering the
emails for "bosse".
2. Can I "move" the db in use for site wide usage and auto-learning?
