*brain-throb* Okay - My system here is the firewall portforwards connections on port 25 to a linux box running exim. The linux box then sends the mail to the exchange server via smtp. Outbound mail goes back the same way.
Since theres no local delivery - can I use spamassassin ? (the reason for the torturous path is to shield the exchange server against the nasty world, and to use some of the anti-spam features of exim as opposed to ms exchange's attitude of "don't hurt me please") On Thu, 2002-05-23 at 15:14, Nick Rout wrote: > > Can you give the list the skinny on how it interacts with the flow of > > mail? > > > > Does spamassassin sit before or after your MTA? or does it bolt on > > similar to a squid redirector? I'm specifically asking about SMTP > > received mail rather than pop/imap/fetchmail. > > Both razor and spamassassin come after your smtp server (not forgetting > of course that this will also catch your fetchmail retireved pop mail > that is fed into your smtp server - at least my fetchmail does anyway. > > On my system both will be inoked from procmail which I have set as my > local delivery agent (ie postfix smtp server feeds to procmail which > delivers the mail, testing it on the way through). I have a > procmail.global file and a procmail.username file for each user. > > procmail.global tests thru razor at present and if it finds a message is > spam dumps it in my SPAM mailbox. If it passes then it goes on to > procmail.nick and gets sorted into various bxes, one for each mailing > list. > > When I get spamassassin working it will also be called from procmail. >