[Exim] Mail loop

Hello,

I've setup two mail servers, with one as a fallback host for the other.
I've also read that in the case of a dns lookup failure, the fallback host
won't get used. From reading the FAQ I spotted a workaround, namely adding
pass_on_timeout to the dnslookup router and adding a manualroute one
underneath. This is what I had

lookuphost:
driver = dnslookup
domains = ! +local_domains
ignore_target_hosts = 127.0.0.0/8
transport = remote_smtp
pass_on_timeout
no_more

dns_timeout:
driver = manualroute
transport = remote_smtp
route_list = * mail-gw2.hursley.ibm.com
no_more

Now, this was all OK except for the fact that incoming mail for domains in
+local_domains was being passed by lookuphost: to dns_timeout: because of
the "! +local_domains" line. When I took that line out, a mail loop
resulted. Could someone please explain why a mail loop should have
happened? I can't see what's wrong.

Many thanks

--
Anjan Sen
System Services, Hursley.
Tel: Internal: 24 7558, External: +44 (0)1962 817558
IBM ASA/ASP/SP certified. RedHat Certified Engineer.