sashk:
> between.?</div><div>5. Their response is:</div><blockquote><div>As
> per the logs so far It your server is making more than one connection
> in less than a second and this is strange as It does not seem to
> be configured this way.</div><div>In this scenario even tho the
If they can't handle RFC 5321 which says try an alternate MX after
a 4xx reply (see below), then you can turn off RFC 5321 retries:
/etc/postfix/main.cf:
smtp_mx_address_limit=1
smtp_mx_session_limit=1
(in addition to setting the _rate_delay so that multiple messages
won't trigger their threshold).
I work for Google now, but I have nothing to with their mail service
(which also has problems with their handling of IPv6 connections).
Here is what RFC 5321 has to say about this:
When the [MX] lookup succeeds, the mapping can result in a list of
alternative delivery addresses rather than a single address, because
of multiple MX records, multihoming, or both. To provide reliable
mail transmission, the SMTP client MUST be able to try (and retry)
each of the relevant addresses in this list in order, until a
delivery attempt succeeds. However, there MAY also be a configurable
limit on the number of alternate addresses that can be tried. In any
case, the SMTP client SHOULD try at least two addresses.
Wietse
