On Wed, 14 Jul 2004, Tony Finch wrote:
> On Wed, 14 Jul 2004, Alan J. Flavell wrote:
> >
> > Seems to me we could adopt a fairly laid-back retry strategy for
> > errors where no MX record is involved, with less-frequent retries, and
> > an earlier final failure than we'd otherwise use? Have folks tried
> > this - any recommendations please?
>
> This is suggested in the Exim spec,
There's an example, yes: perhaps I should have mentioned that - I
wasn't sure how literally to take it.
> and is how we configure Exim here.
* refused_A F,2h,20m;
Just for clarification: you mean that for this type of error you
actually set the final failure at 2hours, and find it to be acceptable
in practice? (Or am I misreading the spec?)
