CUCM 8.6 Upgrade to v10.0

[mohamed helmy]

Hi,

We are planning to upgrade CUCM 8.6.2.24901-1 Cluster (two nodes) to CUCM 10.0
- Backup the running DB v8.6 and Restore it to a transit server with the same version 8.6. 
- Then upgrade this transite server to v10.0
- at the Cut Off, simply make IP Phones and VGW register with the new CUCM v10.0

The question is:
Can we backup DB and restore it to the transit server which has different IP address? or the IP address must be the same?
i.e.
CUCM 8.6 IP : 10.0.0.50
Transite 8.6 IP : 10.0.0.100

also I need to do the same with HA UCCX Cluster.

Regards,

Mohamed Helmy

[Chris Deren]

DRS restore can only be done with the same hostname and IP address, so you will need to perform this in isolation (new server in separate network), restore, then re-ip and change hostname, after which time you can bring it back onto production network.

Alternatively you can look into Prime Deployment to use for the upgrade.

Chris

[Ayodeji Okanlawon]

(+5) to Chris..

In addition may I ask why you are going to 10.0. 10.5 is out and I believe generally a much stable version. I suggest you go to 10.5 and if you have spare VM capacity then use PCD for your migration. It makes your life a little easier

[Jonathan Greenberg]

If you are going to be migrating to a mixed VM environment (ie specs based deployment) where you cannot set up your new environment in isolation. I figured out a way of doing this safely and also make the cut over pretty simple. 

1. Create a new VRF(s) for your new pre-production network.  

2. Create new VLAN(s), associate the VRF and give them the exact same network as your Production CUCM/CUC/UCCX servers

3. Create new port group on VMWare with the new VLAN.

Now you have the same IPs on both environments but they can't see each other. When you are ready to migrate to the new environment, just shutdown both the production/Pre-Production. Change the port-group association of the new servers to the production VLAN and bring the new servers back up. 

You can also use VRFs to route between sites, if you have a multi-site deployment. 

Just keep in mind that for all new UC deployments you need an NTP server, which can be a router. For UCCX, it requires DNS. I typically use a Linux VM running Bind for this purpose. You can change the DNS server IPs to your real DNS once the server is in production. 

 HTH.

    Jon

[Dennis Mink]

Clone your existing 8.6 Virtual machines, then perform your upgrade to 10.

optionally you can also run the "regular backup". 

back out  by restoring your cloned 8.6 machine.