Google Groups no longer supports new Usenet posts or subscriptions. Historical content remains viewable.
Dismiss

When is an SMTP "Delivery Attempt" Successful?

5 views
Skip to first unread message

Philip Gladstone

unread,
Jun 21, 1996, 3:00:00 AM6/21/96
to

SMTP and MX records:

In RFC1123 Section 5.3.4, it says that MX records should be tried
"until a delivery attempt succeeds". However, there is no definition
of a successful delivery attempt. The options would appear to be
as follows:

1) Remote server sends TCP SYN ACK in response to sender's
SYN.

2) Remote server sends any greeting message (including failure
messages).

3) Remote server sends 220 greeting message.

4) Remote server responds happily to HELO message

sendmail-8.7 does option 4.

Can anyone point to a definitive reason why one option should
be picked over another?

Thanks,

Philip
--
Raptor Systems, Inc. Voice: +1-617-487-7700
WWW: http://www.raptor.com/

D. J. Bernstein

unread,
Jun 23, 1996, 3:00:00 AM6/23/96
to

Philip Gladstone <phi...@raptor.com> wrote:
[ when do you give up on the rest of an MX list? ]

> 1) Remote server sends TCP SYN ACK in response to sender's
> SYN.
> 2) Remote server sends any greeting message (including failure
> messages).
> 3) Remote server sends 220 greeting message.
> 4) Remote server responds happily to HELO message

Further possibilities:

5) Remote server responds to DATA.
6) Remote server responds permanently to DATA.

> sendmail-8.7 does option 4.

qmail does option 1.

---Dan

David Collier-Brown

unread,
Jun 24, 1996, 3:00:00 AM6/24/96
to Philip Gladstone

Philip Gladstone wrote:
>
> SMTP and MX records:
>
You asked:

> In RFC1123 Section 5.3.4, it says that MX records should be tried
> "until a delivery attempt succeeds". However, there is no definition
> of a successful delivery attempt. The options would appear to be
> as follows:
>
> 1) Remote server sends TCP SYN ACK in response to sender's
> SYN.
[...]

I suspect that, for a single message, success is the receiveing
mailer responding ``245 OK'' to the data. This is based on
the idea that mail always does a full-fleged handoff, and
never gives up until
a) someone accepts the message, and responsability for delivery
b) a clear refusal is encountered
c) a timeout, measured in days, expires.
In the latetr two cases, the sender is required to send back mail
indicating the delivery could not be done.

Please note that this is difficult to implement when one is batching
multiple mail messages to a site or cloud of sites, and sending to
one or more possible MX targets.
And yes, it's often done wrong.

--dave (once upon a former life I was the postmaster for the
``second largest and second-cheapest'' university in canada) c-b
--
David Collier-Brown, | Always do right. This will gratify some people
185 Ellerslie Ave., | astonish the rest. -- Mark Twain
Willowdale, Ontario | dav...@hobbes.ss.org, unicaat.yorku.ca
N2M 1Y3. 416-223-8968 | http://java.science.yorku.ca/~davecb

0 new messages