800xA 6.0 Upgrade : how to create a new client server network
One of our customers has 800xA 5.1 Rev. E system. All clients, servers and controllers are connected to the same networks (172.16.4.0 and 172.17.4.0).
They want to create a new network for the client / server and keep the 172.16.4.0 network for controllers while upgrading to 6.0.3.
Is this possible? If yes, how would you do it.
Thanks for giving comments quickly as they might decide restart upgrade and keep everything on one network.
Please read below what have been done.
At the end of the upgrade, they would have a new client / server network on 172.16.8.0 and 172.17.8.0. By doing they won't have to shutdown their process to change IP addresses in controllers.
They started their upgrade process with new servers and computers.
They migrated 2008 domain controller to 2012 domain controller. They now have 4 DC running because 2008 DC are combined with aspect servers and must be kept running.
They restored the new system on the new aspect server 6.0.3. It worked. Problems have begun with the next step. The first connectivity server 6.0.3 has been deployed but has trouble to communicate with the aspect server. I'm pretty sure it's related to RNRP. Aspect server has still a NIC connected to 172.16.4 network to keep synch with 2008 DC and connectivity has of course a NIC connected to 172.16.4 network to communicate with controllers. For some reason, they got connectivity services working for 15 minutes (all data from controllers could be displayed on aspect server graphics), but after the next server reboot, connectivity service don't even start like if aspect and connectivity don't see each other.
Yes, you can do it and your Control and Client networks should allways be physically separate - so this is a job you *should* do.
RNRP is just a routing protocol. You can eliminate RNRP issues immediately by simply disconnecting the redundant network. It is more likely that your Network Cards and Computers arent correctly registered on the DNS server any more and that NetBIOS is struggling to resolve the computer names correctly.
- This is not something you should be doing on a live plant. You can break things pretty badly.
- Google the effects of changing the IP address of a DC. This could be causing you trouble. Especially since your new DCs are now on a different network address range from the old DCs. (This is a really good reason to NOT combine the DC and AS by the way)
- Verify that NetBIOS is Disabled on all network cards except the Primary Client/Server network. Check that RSS is turned off, that any power saving on the network cards is off and the the network order is properly set with the Client Server Primary network listed first.
- Verify that the servers are registed on the DNS server **ONLY** with the primary network IP address of the Client server network.
- In the node administration structure check the IP adresses and Names for each server and client.