| 1 |
yes, the @GLOBAL feature in SA's support in MySQL is quite nice. |
| 2 |
|
| 3 |
I've just completed setting up AWL, bayes, and threshold support in MySQL |
| 4 |
for SA. The only thing is I'm using amavisd-new as well so the threshold |
| 5 |
(set here to 5.5) is read from amavisd.conf and not from SQL. However the |
| 6 |
rest is, which is great! |
| 7 |
|
| 8 |
If using amavisd-new and to check for this, run 'amavasd debug-sa' and |
| 9 |
pipe a mail thought it, you will it connects to the db. |
| 10 |
|
| 11 |
Lance(stuNNed) |
| 12 |
|
| 13 |
> Christoph Gysin wrote: |
| 14 |
> |
| 15 |
>> Lance Lassetter wrote: |
| 16 |
>> |
| 17 |
>>> i'm interested if sa-learn can be done site-wide. bayes does seem to |
| 18 |
>>> help a bit but on a system with 500 users it doesn't seem trivial. |
| 19 |
>> |
| 20 |
>> |
| 21 |
>> I have e special IMAP folder, where users can put in false-negatives |
| 22 |
>> (spam that made it in your inbox). A cronjob scans that folder and |
| 23 |
>> trains the sitewide bayes stored in a mysql-DB. |
| 24 |
>> |
| 25 |
>> Christoph |
| 26 |
> |
| 27 |
> I do the same thing. Granted I'm supporting a whopping 3 users, but I |
| 28 |
> use a cron job to go through /home/*/.mail/SA-MissedSpam and |
| 29 |
> /home/*/.mail/SA-FalsePositives nightly. It even clears out the missed |
| 30 |
> spam folder automatically so space doesn't get wasted on what wasn't |
| 31 |
> wanted to begin with. |
| 32 |
> |
| 33 |
> D. Wokan |
| 34 |
> -- |
| 35 |
> gentoo-server@g.o mailing list |
| 36 |
> |
| 37 |
|
| 38 |
|
| 39 |
|
| 40 |
-- |
| 41 |
gentoo-server@g.o mailing list |
Archives