All, I have several remote users in a country well known to inject spam mail. These users are connected via dialup links to our backend Exchange Server and they are able to send email. The Exchange Server relays all email to the front end mail server running Red Hat Enterprise Linux AS 4 and SpamAssassin ver 3.1.7. All incoming and outgoing email is scanned - no ifs, no buts, just is (we've been on a block list due to a compromised machine sending junk email - never again shall we trust all our internal clients)! I have the following in my /etc/mail/spamassassin/local.cf file:
whitelist_from_rcvd [EMAIL PROTECTED] mailsrvr.domain.com Local users send email trigger the ALL_TRUSTED rule but remote users aren't. Some email is even escaping to our clients with a modified subject alluding to the fact that it is spam and hence being relegated to the junk folder by our customers (suffice that to mean that the "no ifs, no buts" clause has been overruled). I can only guess that internal users are firing the ALL_TRUSTED rule as there is only one Received: header (the trusted backend server) by the time spamassassin scans the email. How then can I tell spamassassin to fire the ALL_TRUSTED rule if the connecting host is on the trusted list and ignore further Received: header checks? I have read Mail::SpamAssassin::Conf help file but cannot find the solution yet. I thought the whitelist_allows_relays would help, but I'm not too sure what that does! Is there a better source of documentation somewhere? Cheers ak.
