Showing results for 
Search instead for 
Did you mean: 

Ungraceful Shutdown of UCM Node - Rebuild and Migrate?

Level 1
Level 1
Some time ago, we attempted to upgrade our UCM installation from v11.0 to v11.5. The upgrade process would fail partway through the operation. Also acknowledged that we are way way behind in the upgrade tree.
We were then told by TAC that we were unable to complete the upgrade process because the node had suffered an ungraceful shutdown in the past. The solution previously offered was that we had to rebuild the node entirely, then the upgrade process could be completed successfully.
Currently, our UCM and other related servers (IM&P, Unity, etc.) run on an older ESXI host - it's VMWare ESXI 5.5, accessible through the vSphere client also running version 5.5. We have always wanted to get them moved from that old setup to our other much more modern vSphere 6.7 setup where most of our other active VMs live.
The BIG Question: Is there some sort of solution that would allow us to both:
1. Rebuild the UCM node
2. Migrate it to our vSphere 6.7 setup
We have looked at several approaches previously to get the machines moved from the old host to the new one. We can't use vMotion because we aren't licensed. We can't use Veeam Quick Migration because the old setup is "free" ESXI hosts. 
Are there any other options that we DO have? There must be SOMETHING. Can we create a backup of the old nodes and restore on the new one? Export an OVF template on the old and restore on the new?
Any help or suggestions at all would be so appreciated. And I will gladly provide more details if it will help.
Thank you!
3 Replies 3

Jaime Valencia
Cisco Employee
Cisco Employee

You really should be aware of DRS if you are managing the Cisco UC servers, that has been available for years and does exactly what you want, you make a backup of a cluster, and restore onto a fresh install. Scheduled DRS should be part of any installation contingency plan in case something fails.

The other option if you are migrating + upgrading is going to be to use PCD, you would need to at least upgrade to a higher SU for it to be applicable. PCD migration cannot do the same version for source and target cluster.



if this helps, please rate

As JAVA mentioned, if you are migrating+Upgrading PCD will be the best option. But if its just moving the VM's without upgrading  to another server, PCD will not work.


Response Signature

Level 1
Level 1

We are using DRS to take nightly backups of all our systems, fortunately. 

So if I have this right, we could build the new UCM node on our vSphere 6.7 environment, and then use the backup from DRS to restore there? After which we could upgrade it to 11.5, and subsequently further should we choose to. And we could do the same thing for all our systems as well?

PCD is something I'm not familiar with, but briefly looking at it I do see that it seems like it also would accomplish what we want. I'm going to research that as well.