Aaron Wolfe wrote:
On Tue, Dec 2, 2008 at 3:59 PM, Marc Perkel <[EMAIL PROTECTED]> wrote:
Rick Macdougall wrote:
Marc Perkel wrote:
Thanks Aaron, that is a good point. But I'm running Exim and I think I
can code it so that it will not generate backscatter. I'll have to design
that in up front.
Interesting, how would you do that without dropping email (which is BAD).
Rick
If the recipient is bad then no one would have got the email anyway. But
there wouldn't a a notification to the sender. I suppose I could make it
smarter so that if the message is blessed in one of my many white lists then
I would do a bounce message, otherwise not.
OTOH, if someone is rarely down then the backscatter would probably be
minimal. This will probably be something to experiment with.
no, you will get massive amounts of mail sent to invalid user accounts
regardless of whether the primary is online or not. in fact, when the
primary is online the ratio of mail for invalid vs valid will be
higher. if you accept these messages, you are responsible to send NDR
when the primary rejects them. if you don't do this, you break rfc
compliance. yes, it is usually a waste of everyone's time, but the
one time the CEO of your client's #1 partner mistypes an email address
on a critical message and doesn't get any error back, things go to
crap. be careful with this. consider a cached callout scheme as
mention in an earlier post. it isn't perfect but it is probably good
enough and remains rfc compliant.
If they are online then I do forward callouts to see if the recipient is
valid and based on that I would return a 550 at connect time indicating
an invalid account.