I 've used Linux before in the form of Solaris, Redhat and SuSe, but not in any complex scenarios. It is complex setting up a Linux machine as a mail proxy, and how stable are they?
I've read that Centos 5.2 is a de-labeled version of Redhat enterprise 5, and that I can use Postfix as the mail proxy software. Is it just as simple as getting postfix to listen to the proxy server's NIC on port 25, and forwarding all mail to the Exchange server IP? And then installing SpamAssassin to listen on the port 25 of the proxy?
Main Topics
Browse All Topics
by: robocatPosted on 2008-08-26 at 00:38:29ID: 22312392
It is best to use a front end server that processes SPAM before it gets to your Exchange server, to save resources on your Exchange box.
SA works best on Linux, because this is its native platform. If you are familiar with Linux and installing Linux based software, this is not that difficult to set up.
As far as false positives are concerned, SA calculates a SPAM score for each message. You can tell SA to mark a message if the score is e.g. above 4 and delete the message if score is e.g. above 10. Messages with a score between 4 and 10 are delivered normally, but the user can automatically filter this out to a SPAM mailbox for manual review.
We almost never get false positives with a score above 10, but if this happens, you could use a whitelist. Maintaining blacklists/whitelists requires editing text files.