This is hard to
explain and also be short - I'll try. My mail server has 2 IP addresses
and a call to tech support suggested that any mail directed to either IP will be
picked up and processed. If a blacklist is on one IP (.1) and not the
other (.2), mail can be directed inbound (to .2) and bypass the blacklist
on the first (.1).
Clearly there would
never be the case where email is sent to an IP that is not the listed MX.
Alas this no longer seems to be a true statement (if it ever were
<grin>).
The dilemma: I
have a blacklist that says this small range of IPs (say Postini
or some services source addresses) are the only ones not in the DNS
data base blacklist I created. This was to be used as a drop everything
trusted blacklist for it was my own list and and geared to a known
service. I could not simply filter on the router all traffic for I have
many many virtual hosts. I wanted to have custom blacklists per
domain. Then learnt I needed these per IP, then learn that "misdirected
mail" would be picked up on any IP and there seems to be no way to prevent
this.
I need to ask the
list for ideas on shutting down what I call "the side door" with the
understanding that every mail domain has unique needs.
Ian McCutcheon
PS: an
aside, if anyone can supply me with the information needed to understand the
correlation between an IP address and a virtual domain I'd find that info
enlightening. I was only able to assure myself that the registry told me
that there were IP addresses (official hosts) and virtual domains but no way to
see which virtual was more closely associated with which IP
address.
Last:
Anyone have their user's web mail defined rules go wonky after 8.11
patch?
