Failing Over a Replica with Veeam B&R v12

In the previous post, we saw how to create a replica in Veeam B&R v12.
Now I’ll briefly show you how to bring that replica into production if needed, and we’ll discuss a few use cases.

Let’s go!

In the Veeam B&R v12 console, we can see the replication job.

Under Replicas, we have one ready to use or start.

Right-click the replica you want, and you’ll see these options:

Let’s briefly explain the two features I use most.
Failover now ->: Veeam starts the replica on the host we configured. This is the usual option if production is lost or stopped.
Planned failover ->: Safely shuts down the source and starts the replica without data loss, although there is an outage while the source powers off and the replica powers on.
You can use it to test replicas or run a VM on another host if you don’t have vMotion or something similar.

In our example, we’ll simulate a production outage, so choose “Failover now” and continue.

Here we can see the replica VM, Photon_ova_4.0_replica, is powered on.

And we can reach the machine with a PING!

Once the machine is running as a replica, these are the final options:
Permanent failover: the replica remains as the VM.
Undo failover: discard all changes made to the replica and power it off.
Failback to production: copy changes made to the replica back to the production VM, then power off the replica and start the “original.”

If we lost the source machine or its storage, Permanent Failover is the better option. If we only lost connectivity to the storage array, we can use Failback to Production once connectivity is restored, copy over the changes, and continue as usual.

Let’s run Failback to Production to restore the original setup, copy changes from the replica back to production, and resume normal operations.

NOTE: I selected the option to power on the destination after the data transfer.

Here we can see the transition: the original machine was unreachable and is now back online.

As a final step, after failing back the replica to production, we need to commit the operation. This completes the process, applies all changes, and returns the machines to their production and replica states.

Regards.