Re: [Exim] temporarily rejected RCPT -> fixed itself?

Hi Marilyn, on Mon, 17 Nov 2003 12:29:14 -0800 (PST) you wrote:

> We accidentally pulled the plug on our mail-server and when it came
> back up, exim 4.22 was temporarily rejecting all incoming external
> mail, looking like this in the log:
> So, what happened?

Bit of a guess this, but it looks like it could well be to do with DNS.
How is your DNS handled - do you have a resolver (e.g. BIND) running on
the same machine as Exim? If so, I've seen the same thing happen a number
of times.

You'll have to excuse me being a bit vague here, but I've never had time
to look into the details of the configuration or how BIND works to find
out precisely what's going on, but basically I've seen a number of times
that BIND (at least in some default installations of Red Hat) isn't happy
(i.e. runs but doesn't resolve anything) in the immediate aftermath of a
reboot, which consequently causes issues like you saw (the temp failure
being due to an inability to resolve the sender's domain as you have
verify=sender).

The simple fix is just to restart BIND and things start working (in fact,
I think I recall doing a "hack" on some machines which HUPs BIND shortly
after reboot just to make sure this doesn't happen). I would speculate
that it's some kind of timing issue, where the network interface(s) aren't
fully initialised before BIND starts or something along those lines.

Is that any help?

Tim