I run a mail server that accepts mail for a couple hundred domains, and we 
utilize qmail heavily.  We are using qmail-scanner to virus scan some of
the domains, as well as spam assassin.  I'm in the process of integrating
SA via qmail-admin (right now it's being run out of .qmail-default).

I've been thinking about ways to reduce the amount of spam making it into
our network.  When qmail sessions are established, you usually get the
remote server's IP address in the environmental variable $TCPREMOTEIP or 
$TCPREMOTEADDRESS (the variable name escapes me at the moment).

While I have absolutely no intentions of dropping/deleting anyone's mail
based on score, is there an easy way to modify Spam Assassin to log the
IP address of a "hit" individually?

My theory here is if I could log the IP address from that environmental
variable, I could keep statistics on hosts that send spam.  If I could
integrate this into a MySQL table I could time stamp the entries and make
the ip address field unique.  A clean up command could delete any entries
that havn't been updated in the past X days, and a few select statements
could show you every host that has "hit" more than X times.

It'd be a good way to keep track of remote hosts that tend to send spam, 
and it could even be setup so that if a host IS in the table for being
a spammer, we could also measure how much "legitimate" mail comes from
that domain as well.

Then, someone could review the data and decided if they wish to internally
block that machine.  I am confident there are people out there who are
preventing them selves from entering blacklists, but give themselves
permission to relay through their own equipment to spam.  This would
help crack down on that to some degree.

I imagine someone may have already begun something along these lines, so
I wanted to bring it up and suggest it before I start hacking at our
spamc/spamd implementation.  Unfortunately, in the spamc/spamd scenario,
the environmental variable won't be passed to the perl daemon.  

However, maybe this could be done inside qmail-scanner.  Maybe spamc 
exits with a specific error code if it gets a positive result from spamd?

That might be the answer to my project..

-Jeff


-------------------------------------------------------
This sf.net email is sponsored by: Influence the future 
of Java(TM) technology. Join the Java Community 
Process(SM) (JCP(SM)) program now. 
http://ads.sourceforge.net/cgi-bin/redirect.pl?sunm0004en
_______________________________________________
Spamassassin-talk mailing list
[EMAIL PROTECTED]
https://lists.sourceforge.net/lists/listinfo/spamassassin-talk

Reply via email to