Bowie Bailey wrote:
LDB wrote:
I am invoking spamc through a filter script where spamd is
listening. Also, I am using PostFix as the MTA. My platform is
Debian Linux. I am SA version 3.0.3.
The below config. captures about 1700 spams a day but it is NOT
enough.
Can anyone kindly suggest a better more strict config. for me or
just simply critique the below config.?
########################################################
lock_method flock
#all_spam_to [EMAIL PROTECTED]
required_hits 5
spamcop_from_address [EMAIL PROTECTED]
#spamcop_to_address [EMAIL PROTECTED]
use_dcc 1
dcc_path /usr/bin/dccproc
dcc_dccifd_path /usr/sbin/dccifd
use_razor2 1
use_pyzor 1
user_bayes 1
use_bayes_rules 1
bayes_auto_learn 1
report_safe 1
rewrite_header From [EMAIL PROTECTED]
rewrite_header To [EMAIL PROTECTED]
rewrite_header Subject [****** SPAM_SUSPECTED ******]
ok_languages en es fr de it ja zh
bayes_path /var/spool/spamassassin/bayes
#########################################################
As Rick already mentioned, upgrade and make sure the Bayes DB is
properly trained.
Also, I notice that you don't have a trusted_networks setting in there.
If you don't specify it, SA will guess and may get it wrong. This
will throw off quite a few of the header and network checks. I would
recommend manually specifying this setting even if it looks like it is
working properly without it.
You can also post the headers from one or two of the spams that are
getting through and we may be able to make some more suggestions based
on that.
Thank you Bowie ...
LDB
