ram a écrit : > On my MX servers we accept the mail and relay it to a spamassassin > server > > Now for someids we dont spam-scan the mail ( for eg ab...@domain.com ) > If a spammer marks a mail to some real recipient and to ab...@domain.com > the mail goes thru because any mail for abuse@ is not scanned > > > Can I configure postfix on the MX server to send mails for > ab...@domain.com in a seperate transaction and other recipients in a > seperate transaction > I dont want to break all the multi-recipient mails into multiple, only > for those mails marked to ab...@... > >
here are some choices: - use multiple postfix instances, and use transport_maps instead of content_filter/FILTER (you need multiple instance because transports are global) - use a filter that supports per recipient policies. amavsid-new and dspam can do that - pass all mail through the filter, but ignore the results for ab...@*. This is what I do, because this way I get spamassassin infos in the headers, which comes in handy if it's a fake complaint.
