On Thu, 2010-01-14 at 21:55 +1100, Jean-Yves Avenard wrote: > How can I write a rule on the primary server that will automatically > consider the message as spam is the other server detected it as spam. > > I tried: > header PREVIOUS_SPAM X-Spam-Flag =~ /YES/, > header PREVIOUS_SPAM X-Spam-Status ~= /Yes/ > etc. > > None of them seems to make any differences.
I'd go for marking with different headers on either of the servers. And, depending on the apropriate approach, just pass the mail already marked as spam by the first server to user (or his spam folder) or feed it to SA on the second machine but with scoring based on the header from the first server. BTW, as the check definitions seem to be pretty ok at first glance, did you set scoring for those rules, or did you just add the rules? You should get something like: header PREVIOUS_SERVER_SPAM X-Some-Unique-ID-Spam =~ /\bYes\b/i describe PREVIOUS_SERVER_SPAM Already scored score PREVIOUS_SERVER_SPAM 2000 -- /\-\/\-\/\-\/\-\/\-\/\-\/\ \ k...@epsilon.eu.org / / http://epsilon.eu.org/ \ \/-/\/-/\/-/\/-/\/-/\/-/\/
