Liam:
> Thanks for your help in sorting this out, Wietse.
>
> On Sun, Nov 7, 2010 at 11:19 AM, Wietse Venema <wie...@porcupine.org> wrote:
>
> > > I can support per-node domains, e.g. node-id.mailhost.xyz
> > I suppose some wild-card DNS record will do the trick.
> >
>
> So that all those node domains resolve MX to my proxy?
You mean A records. LMTP does not use MX.
> The non-optimized slow path is:
> > incoming queue -> active queue -> deferred queue
> > [try to deliver while the destination is down, with exponential
> > back-off]
> > deferred queue -> active queue -> deferred queue + append fast-flush
> > file
> > deferred queue -> active queue -> deferred queue + append fast-flush
> > file
> > ...
> >
>
> But surely I can prevent most (lmtp-based) retries to my proxy with a config
> setting? And that should keep the load for this 15% manageable? Are there
> other optimizations I can apply?
You will need to maintain some low retry rate, otherwise mail will
accumulate forever for dead clients.
I'm stepping out of this thread now, because I need to focus on
other stuff. Someone else can continue this thread if they like.
Wietse
