From: Tony T. <Ton...@go...> - 2015-06-02 22:30:45
|
We have been using Pyzor with SpamAssassin and Postfix, and I notice that a lot of the spam that gets through 'untagged' is correctly detected if I manually re-scan them. I'm assuming the difference is the time it takes for me to notice and test those spam. I also notice that we get a lot of duplicate spam to current employees (mostly senior mgmt/owners) as well as long-gone employees. So I would really like to setup an additional local Pyzor server and train it with email from the long-gone employees. The goal is to immediately identify spam rather than waiting on the public servers to update. But I have not been able to find any HOWTO's for such a setup. Can someone please help me? I think the general procedure would be: 1. Install pyzor server 2. Create rule in Postfix or SA to redirect known-bad messages (based on old-old-old addresses) to a separate instance of Pyzor client 3. This second copy of Pyzor client would be configured to 'report' all messages it receives to the local Pyzor server 4. Update the existing primary Pyzor client to perform lookups against the local Pyzor server, either in addition to or instead of the public. Unfortunately, I really only know how to do the first step. - Tony |