At 04:44 PM 5/03/01 -0400, James M Galvin wrote:
>One thing you might do is look in your connection log files for
>additional information. A problem I have seen before (many times
>unfortunately) is the other side having "tuned" the timeouts for one or
>more of the transactions in the SMTP dialogue, which usually results in
>them either giving up too soon or not soon enough. <SNIP>
After sending my post through, that exact thought came to
mind. It's the most logical reason. Unfortunately, I can't
really tell more from the logs than what I've posted. Our
logs seem to take everything from hello to goodbye and
closing the connection.
>As far as what you could do about it at your end, is it safe to assume
>you don't actually look at the "bounce" message, which suggests you've
>got the actual failed address coded in an envelope address? If so, you
>might consider how you could get your 'bounce' processing to look at the
>message so it could ignore autoresponder messages. My system actually
>looks at the subject line and the first three lines of the bounce
>message explicitly for exactly this reason.
I'm not sure how this works. I'm a non-techy (scary isn't
it? but I seem to do better than most mail admin people LOL)
and I do know that copies of bounces do go to the list owner
unless they choose to suppress them, but autoresponder messages
don't. (The problems we're having are unique to the servers
we're running Post.office on.) The autoresponder messages
that are getting kicked back by senate.gov addresses have
a different subject line (no original message contents are
repeated) and there is no text in the body quoting back
the original message. Logically, these bounces SHOULD be
going direct to the FROM address or REPLY TO address in
posts, but they are actually going to the return-path
address instead.
Sharon
