ESX Server Hardware replacement - CUCM, CUC CUPS

[russjstewart]

Hi,

I am hoping someone with real world experience can cast some light on this for me please.

My problem is that the existing Cisco servers that host our VM servers for CUCM, CUC and CUPS/IMPS have reached end of life and need to be replaced. We are on 12.0.1 and do not wish to upgrade as part of this process.

 Each physical server hosts 3 VM with the publishers on one host and the subscribers on the other

TAC have told us there is a problem with using the PCD tool will not work as there is a bug when using it to move the same version (defect CSCvg33737 )

So we are looking creating the VM, backing up the existing VM using DRS then shutting the existing VM restoring onto the new VM.

I am hoping someone has done this and can give me some tips or warnings on things to look out for.

Thanks for your time

[Nithin Eluvathingal]

PCD is preferred than  traditional migrations.But Unity migrations are not still possible with PCD. for unity  you need to do the traditional method.  since the bug is only for migrating to same version, why don't you try next version for rest of the applications. 

• make sure that you have the security password for all the server which you are going to take DRS backup and restore.
• while taking the backup you may get stuck with the DRS Master and local service. in some  cases I have restarted these services for the backup to work.

you can go through below link and it covers all prerequisite for backup and restore.

https://www.cisco.com/c/en/us/td/docs/voice_ip_comm/cucm/admin/12_0_1/admin/cucm_b_administration-guide-1201/cucm_b_administration-guide-1201_chapter_01011.html

Restore Prerequisites

• Make sure that you meet the version requirements:

• All Cisco Unified Communications Manager cluster nodes must be running the same version of the Cisco Unified Communications Manager application.

• All IM and Presence Service cluster nodes must be running the same version of the IM and Presence Service application.

• The version saved in the backup file must match the version that is running on the cluster nodes.

The entire version string must match. For example, if the IM and Presence database publisher node is at version 11.5.1.10000-1, then all IM and Presence subscriber nodes must be 11.5.1.10000-1, and the backup file must also be must be 11.5.1.10000-1. If you try to restore the system from a backup file that does not match the current version, the restore will fail.

• Make sure that the IP address, hostname, DNS configuration and deployment type for the server matches the IP address, hostname, DNS configuration and deployment type that are stored on the backup file.

• If you have changed the cluster security password since the backup was run, make sure that you have a record of the old password, or the restore will fail.

[russjstewart]

Hi NithinElu
Thanks for the information. I will discuss with the server team and who are doing the build

[russjstewart]

Hi Nithin,

Sorry for not getting back earlier, but with COVID things have been a bit chaotic.

In the end the server team configured ESXi and powered down the VM's did a VMware Converter to clone the powered off VM's to the new hardware. The hardware was not part of our VCenter.

All VM appear to have come across cleanly even Unity. DB replication is working, licensing is up to date.

Thanks for taking the time to reply and for your input, it was greatly appreciated

[Roger Kallberg · Accepted solution]

No need to do all that complex steps. Simply install the new hosts with ESXi, add them to vcenter and move the current VMs for your UC environment over. If all is setup correctly it would just be a rehost vmotion. If you’re not using SAN storage that all hosts can access or at all, it would be a vmotion with rehost and change of storage.

The offline option would be to shutdown the VM and copy the VM files across from the old host to the new.

[russjstewart]

Thank for the info, will talk to the server team who are doing the build

[russjstewart]

Thanks for the information. I will discuss with the server team and who are doing the build

[russjstewart]

Hi Roger,

Sorry for not getting back earlier, but with COVID things have been a bit chaotic.

In the end the server team configured ESXi and powered down the VM's did a VMware Converter to clone the powered off VM's to the new hardware. The hardware was not part of our VCenter.

All VM appear to have come across cleanly even Unity. DB replication is working, licensing is up to date.

Thanks for taking the time to reply and for your input, it was greatly appreciated

[VIVIEN MAHONEY]

We used Roger's suggested method recently to migrate our 12.5 infrastructure to a new physical system.  We were migrating from standalone C240 to B200 Blades with network storage.  We ran final backup, shut down the affected VMs, copied the VM storage to shared network storage, then Vmotion the (powered down) VM nodes.  After starting on new servers, did full validation. 

No issues except the time to copy node storage was a bit longer than we initially estimated, particularly for the Unity Connection system.