Re: [exim] Strange log entry about backup MX

Interesting discussion, this.

For a given well-designed MX platform where the MX servers are
physically separated and are on different networks with different
upstream providers, there are myriad reasons why any two MX servers for
a given domain might not be able to reach each other. That means they
might see each other as "down" when in fact they're reachable from
everywhere but themselves.

There are also myriad reasons why a given client - that is, an
"external" MTA trying to deliver a message - might only be able to see
one of those two. In the case where the two MX servers can reach each
other, but a client cannot, delivery delays could occur.

In the more condensed scenario, where there are two MX servers on the
same network which share the same upstream there are still myriad
reasons why a given external client might only be able to reach one and
not another.

In the end, however, none of this matters - as long as the MX returns a
4xx response (that is, temporary) then the sending MTA should requeue
and retry - and will probably deliver the message later.

To go back to the OP's question: you need to check your logs to work out
why Exim isn't speaking to the primary MX.

Graeme