We have 1 master 3.8sp4ir3 bordermanager server running on netware 6.5, and
several slaves all over the world.
One slave is located in the USA. The tunnel however at this moment is not
used and all emailtraffic from the USA will be directly delivered from their
exchange environment to the other domains. If they sent email from their
domain to our domain, then it is succesfully delivered to our mailserver
(the incoming email to our domain is delivered to our public nic of the vpn
master server). However when we stop the vpn at our master server, the email
to our domain stays stuck in the queue on the exchange server.

The bordermanager slave there in the USA is configured with a public
ipaddress in the private address range. Nat is performed on the router.
The master and all other slaves do use real public ipaddresses.

Anyone experienced this before? the domain mx record can be succesfully

Other strange thing is that they configured a secondary public ip that
connects to a terminal server, which we can not connect to from our site
(except by using vpn). If I connect from my homeaddress it is possible to