Hi, Just installed SA on a postfix system. I also use Razor. I have 4.0 as a
threshold.
I usually get around 100 spam a day. it used to take me 15-30s to delete them
but lately I couldn't delete them everyday and that makes it a bit more
complicated to find the ham.
I've looked at the FAQ but I might have missed something.
Thanks to SA devs. (if they are listening here). It takes around 40% of my
spam.
Here are my questions:
- 40 % of the spam is that what I should expect? with razor helping?
- If I understand it right, the files in "/usr/share/spamassassin/" are SA
default rules. Among those rules, Some find the word viagra. I've received
mails with "viagra" (oops, I hope this list is not using such a filter) that
where not marked as spam. Hmm after checking I see that the test is /VIAGRA/
all uppercase. Shouldn't that one be case insensitive in the distribution?
- in X-Spam-Status, I see "autolearn=no". Hmm, autolearn what? and shouldn't
that be 'yes'. A pointer to some doc would be great.
- Running SA manually on a mail gets a much higher score. The mail goes
through SA automatically and I get the following score:
X-Spam-Status: No, hits=3.5 required=4.0 tests=NO_COST,RAZOR2_CF_RANGE_51_100,
RAZOR2_CHECK autolearn=no version=2.63
but when I run SA one more time (without removing the SA tags) I get this:
X-Spam-Status: Yes, hits=9.4 required=4.0 tests=NO_COST,
RAZOR2_CF_RANGE_51_100,RAZOR2_CHECK,RCVD_IN_BL_SPAMCOP_NET,
RCVD_IN_DSBL,RCVD_IN_DYNABLOCK,RCVD_IN_SORBS autolearn=no version=2.63
Which gives?
- user filter should be anabled in /etc/mail/spamassassin/local.cf and then
added in ~/.spamassassin/user_prefs. Right? (I'll start adding rules if the
above can't be fixed)
Cheers, Nadim.
