The restore process for a TKG cluster is manually triggered through the PowerProtect Data Manager UI. The process is similar to the backup process.
Steps 1, 2: CNDM interacts with VM direct to find and reserve vProxy, which is pre-installed and is responsible for moving data from the backup target to FCD. CNDM then initiates communication with the API server of the guest cluster by sending RestoreJobCR that contains the restore request information. It also passes details to the API server about which vProxy is being used to restore the data.
Step 3, 4, 5, 6: The API server of the guest cluster provides the PowerProtect controller with RestoreJob information, and the PowerProtect controller creates PVC and then Velero RestoreCR. It informs the API server about PVC and CR created. The API server then communicates with Velero PodVM by sending Restore CR to the vSphere plug-in within the Velero pod.
Step 7, 8, 9: Velero PodVM restores the metadata from the backup target and communicates with the PowerProtect controller. The PowerProtect controller asks the vProxy VM to restore the data using the backup driver. Once the restore is completed, the guest cluster is notified about the job completion.