Hi Fellow dspam users,
I am new to the list.
This week I have implemented dspam successfully with postfix using transport maps.
In less than 48 hours dspam is filtering out those difficult spam mails which passes
the postscreen tests.
Hereby I share my dspam implementation strategy and choices.
I have a basic setup. I use postfix for all the selection logic. With postfix I can
filter which (sub)domain or e-mail address will be passed to the dspam socket.
Dspam reinjects the mail into postfix and postfix depending on the (sub)domain
or user, sends the mail to a particular internal protected mailserver.
Here are relevant snippets of my dspam configuration file.
# keep it simple
# reinject into postfix using smtp
# for now train on everything
Algorithm graham burton
# as advised
# tag the mail with a specific message.
# using the no spam tag, a companywide disclaimer
# can be enforced (just a matter of creativity)
# again keep it simple
# make a postfix transport to retrain dspam
# User@domain is equal to user@Domain
# make dspam talk lmtp
# spam mails are delivered after being tagged and after
# the subject has been preceded with [SPAM]
# On the internal mail server a companywide rule tags
# the mail as junk and move it to the users SPAM
ServerParameters "--deliver=innocent,spam -d %u"
---- end relevant parts of dspam config
# group file content
# global-user preferences file content
Users don’t want to be bothered with details. They want
a configuration which works out of the box.
Optionally users can send a mail to [hidden email] or
[hidden email] to retrain for their specific account.
The system is fed by a couple of motivated users and because
of the group type everyone will benefit.