R 0
meta HABEAS_ACCREDITED_COI 0
meta HABEAS_ACCREDITED_SOI 0
meta HABEAS_CHECKED 0
Roger Marquis
(up to 4 or so, to avoid DNS timeouts). Postfix supports
this with "warn_if_reject" before doing the actual "5XX" reject. It's the
warnings that yield valid data, or at least they do with large and
representative samples (which IME >= 100K msgs/day).
Roger Marquis
LuKreme wrote:
PostConf http://www.postconf.com for example.
Looks interesting, but not FreBSD demo :/
Waiting only for a postfix port with an "overwrites-base" option.
The code itself works with any postfix home directory.
Roger Marquis
x27;t that complex, and some web front-ends implement them, for
RBLs and RHSBLs, in block or warn-only modes, and generate reports based on
the results. PostConf http://www.postconf.com for example.
Roger Marquis
uld probably take a few
hours to complete. Wish I had the time to do it for free.
OTOH, if you have scripting skills it's pretty straightforward. The only
time consuming part is parsing the data down to just the essential elements
and formatting it for easy reading. A few dozen iterations and voila.
Roger Marquis
This is not correct. AWL is *not* taken into account for auto-learning.
Did not mean to imply that it was. Apologies for the lack of clarity.
Most of our negative experience has been with autolearn. Based on your
post we are retesting AWL starting with a low threshold (0.1).
Roger Marquis
absolutely essential for monitoring production Unix/Linux
systems.
Roger Marquis
's) Bayes DBs become corrupt from autolearn
/ AWL...
IME you'll have better results by manually processing only a small amount
of representative ham and spam through sa-learn. By small I do mean only a
few per week.
Our Bayes Tips are online at: http://www.postconf.com/docs/bayes_tips.shtml
Roger Marquis
