I'm worried about referrer spam. The situation I have is my users send
their email to smtp1.mydomain.com, which checks for mime errors then
relays the mail to smtp2.mydomain.com which scans for spam and viruses.
My problem is that if smtp2 refuses a message because it's a virus or
spam, then smtp1 will send the bounce message to whatever the address
that the malware sent the mail as.
The best way to sort this out would be to allow each IP to send from
only certain domains. Apart from this being a major problem to
implement for 5000 users, we are using dynamic IP dialups so this
scenario becomes close to impossible.
Is there any other way of solving this problem, or will I have to move
the spam and virus scanning onto smtp1?
