----- Original Message -----
To: "Hector Santos" <hsantos(_at_)santronics(_dot_)com>
We eventually settled on an intermediate strategy. Temporary failures
before the MAIL FROM cause us to retry using the next MX, failures
at or after the MAIL FROM do not. We have found that this works
pretty well overall.
Same here Ned. Quickly looking at the current next MX logic:
- Connection errors,
- non-220 response at connection
- non-45x or non-55x responses (i.e., like 42x) at other states
- and drops.
We track the bad connections(host) and if its reaches a certain
point, then the transaction is stopped and blocked. Future messages
for the same host have a "lets try again 1 time logic" otherwise
it is permanently blocked until cleared by the admin.
We don't do the permanent admin block (at least not automatically),
but we do all the rest.
Admittedly, its one of those "Did you check your auto bad domain or bad
address list?" support question answers. :-) But no doubt, the pros
outweigh the little reports. All optional of course, we have it tied to our
RCPT TO lookup response too. So when the system tracks unreachable host or
email addresses, the next time the RCPT TO is attempted with our mail bots,
it returns a negative response. e.g., it helps with the list server
integration during a SMTP distribution for auto inactivating subscribers.
Most complaints are typically from the same people who never seem to keep a
stable server. :-) And its not even complaints, they just ask to be removed
from this list like they are already accustom to the idea they can't ask for
too much when they know they can't keep a server running. :-)
Hector Santos, Santronics Software, Inc.