Author: Alan J. Flavell Date: To: Exim users list Subject: Re: [Exim] Departed users, forwarding addresses and collateral spam
On Wed, 21 Jul 2004, Kjetil Torgrim Homme wrote:
> On Sat, 2004-05-29 at 13:40 +0100, Alan J. Flavell wrote:
> > I'd like to raise the idea that, where such forwarded addresses no
> > longer ever send mail, it would be good practice for mailers to only
> > accept actual mail items, i.e items with a non-null envelope sender,
> > when they are addressed to one of these old addresses; and should
> > reject anything with a null envelope sender, on the grounds that there
> > can be no valid delivery status notifications in relation to an
> > address which no longer ever sends mail.
>
> interesting idea, but I don't think this policy could or should be
> applied generally.
Confirm that this was intended only for addresses which are no longer
in use as sender addresses, and are accepted (and forwarded) here only
for the convenience of departed users. The number of such addresses
is relatively limited - in our situation we're talking about a few
dozen at most, not thousands...
> it would be hard to write the ACL, too -- you don't
> want this to happen if the user has more than one forward target, and
> one of them is his local account.
The users which I have in mind here, no longer possess a local
account; they are only entries in an alias file.
> I know we have many alumni who use the university address as their
> invariant address as they move from job to job,
Yes, I understand that requirement, but it's not that kind of user
which I have in mind here. Typically, at some point in the next year
or so, each of these addresses will be phased-out entirely, and this
forwarding alias is only a courtesy for their convenience during the
transition. Unfortunately, some of these addresses are getting
enthusiastically faked as senders of spam and/or viruses, and we then
get all the stupid bounces (collateral) from sites which have
misguidedly accepted the spam or virus, and then decided they didn't
want it after all, and are trying to report the problem to the faked
sender address.
> so I think it would have to be configurable on a per-recipient basis. .. > and then you might as well implement return-path rewriting, and STRONGLY
> suggest that people who's only forwarding and never use the e-mail
> address for new e-mail, turn it on.
These users no longer have an account on our system, so any changes
to the alias have to be made on their behalf by the mail admin.
There's nothing wrong with what you're saying - it's just that you're
addressing a significantly different scenario than the one we're
dealing with.
