On Sat, 10 Aug 2002, Chris Knipe wrote:
>
> > If the first server goes down, mailservers from the net will contact
> > your secondary MX, it handles as much as it can and queues mails for the
> > first server until it can be reached again. It will recognize that the
> > first server is down simply by being unable to reach it. Where's the
> > problem?
>
> That's how I understand it as well, and to my understanding how it should
> work. However, what is required on the secondary MX machine to let it know
> to queue and re-route the domains?
Nothing. As long as it will accept mail for the domain (relay_domains or
relay_domains_include_local_mx for exim3, or appropriate acl's for
exim4), and it knows how to route them to the primary, it will accept
them, and then when it finds it cant reach the primary server it will
defer, and retry according to your retry rules. Your probably want to
set a fairly high maximium retry period for these domains (at least as
long or longer than you ever expect the primary to be down.. I wouldnt
go less than 7 days, more is ok)
>
> A domain for which the secondary MX is there, the domain cannot be specified
> in local_domains because the secondary MX will not process any deliveries
> for the domain, so what exactly is needed here?
>
> The manual "re-route" I get and understand, I can process it with a simple
> router, ex:
> backup_MX:
> domains = backupdomain.com
> driver = domainlist
> transport = remote_smtp
> route_list = * mailserver.backupdomain.com byname
> self = fail_soft
>
> I'm sure I'll need something else to tell exim to accept messages for
> backupdomain.com however, because it is not specified in local_domains
> (which it cannot be). Would this be a entry for relay_domains on the
> secondary servers then?
>
>
>
>
> --
>
> ## List details at http://www.exim.org/mailman/listinfo/exim-users Exim details at http://www.exim.org/ ##
>
>
