Multiserver deployment for Data protection for VMware

Sid

Newcomer
Joined
Jun 29, 2021
Messages
1
Reaction score
0
Points
0
Hi,

We currently have an SP server running on the AIX platform, with Data Protection for VMware protecting around 70 VMs. Recently, we've deployed a new SP server based on RHEL and are looking to implement Data Protection for VMware from scratch in this new environment.

Our strategy involves a multi-server deployment, leveraging IBM documentation link below to configure it with the current vCenter plugin as additional backup server, along with existing default backup server. However, we have a challenge as we can not remove the default backup server from the plugin, and decommissioning the old (default) backup server is not possible.



I have an alternate plan to deploy a new VBS (New Primary Data Mover) without modifying/registering the existing plugin. We'll then set up backups using the VMware vSphere GUI on the New Primary Data mover, gradually phasing in VMs. Eventually, we'll decommission the old primary data mover/plugin. Following decommissioning, we can then invoke a configuration wizard to install the new plugin with the New Primary Data Mover.

My query is related to the validity of deploying dual vStorage Backup Servers (Primary Data Mover), each registered with separate SP servers for the same vCenter?

I could not find any high availability design documentation to have dual VBS (Primary Data movers). Any suggestions or guidance on this approach would be highly appreciated.

Regards,
Sid
 
My query is related to the validity of deploying dual vStorage Backup Servers (Primary Data Mover), each registered with separate SP servers for the same vCenter?

I could not find any high availability design documentation to have dual VBS (Primary Data movers). Any suggestions or guidance on this approach would be highly appreciated.

If I recall correctly, I had asked IBM about HA years ago and the answer is it is not supported.

Reason:

For HA to work correctly, the application itself must have been designed to work in a HA mode. It should recognize Network, Disk and other failovers.

Having said this, there no one stopping you to try doing this by:

- creating a HA environment at the OS level - search the Web for examples; there are tons of this
- install the application binaries on the common disk (caveat: this will be your common failure point)
- the network and other resources are failover points and should be transparent to the application
- the network has to be configured in failover mode

Try this approach if you have the time and hardware to work with.

Good luck
 
Last edited:
Back
Top