I haven't delved too deeply into the use of spamassassin, yet.
Please feel free to correct any assumptions I may be stating here. It
appears that with the plugin, one can currently drop the connection with a
5xx error (return DENY). I presume we could return a DENYSOFT to give a
4xx error instead. I am also presuming that if we set the
munge_subject_threshold value higher than the reject_threshold, we would
ensure that no munging ever takes place.
I have to main concerns with implementing spamassassin. The first
is the processing required. Can anyone offer their impressions about how
much overhead it takes and how this could affect a really busy server? My
own server is not that busy, but we sometimes build mail servers for
others.
The second concern is the likelyhood of bouncing or dropping legit
commercial email. Things like notices from Amazon, any software companies
you've registered with, etc. Is this just a matter of determining which
spamassassin checks to run and what threshold score to use (i.e.
tweaking)?
Thanks for any feedback.
We wish you a Merry Christmas and a spam-free New Year.
--
Roger Walker
"His Pain - Our Gain"
