[Mimedefang] Powerful anti-spam rule that never discards vali d e-mail

[Les Mikesell]

> From: Mycrom
> This all is a good idea but you guys need to change the thinking a little.
> Issuing a 4XX to all messages will cause the message to be delayed in most
> cases. The remote mta will take it and stick it into its little pocket until
> next queue run. What you should do is a connection reset or just drop the
> STMP session at some point. This is treated differently and should make most
> normal MTA's just hop over to the next priority MX. This again would be
> using the same principal that most spamming tools would just hiccup and not
> act as a proper MTA.

It may not be this easy.  I just did an overnight run with a higher and
lower MX record pointing to a firewalled address that would return an
ICMP DENIED instead of connecting and still saw about the usual amount
of spam coming in so most of the senders must do the quick retry to
an alternate MX.  I didn't do anything to log the rejections so a few
might have given up but it wasn't dramatic.

--
  Les Mikesell
    les at futuresource.com