You can backup and restore workloads running on TKG clusters on Supervisor using the Velero Plugin for vSphere.

Prerequisites

To backup and restore TKG clusters workloads using the Velero Plugin for vSphere, you must first install the Velero Plugin for vSphere on the target cluster. See Install and Configure the Velero Plugin for vSphere on a TKG Cluster.

Backup a Workload

Below is an example command for creating a Velero backup. 
velero backup create <backup name> --include-namespaces=my-namespace

Velero backup will be marked as Completed after all local snapshots have been taken and Kubernetes metadata, except volume snapshots, has been uploaded to the object store. At this point, asynchronous data movement tasks, i.e., the upload of volume snapshot, are still happening in the background and may take some time to complete. We can check the status of volume snapshot by monitoring Snapshot Custom Resources (CRs).

Snapshots

Snapshots are used for backing up persistent volumes. For each volume snapshot, a Snapshot CR is created in the same namespace as the persistent volume claim (PVC) that is snapshotted.

You can get all Snapshots in PVC namespace by running the following command. 
kubectl get -n <pvc namespace> snapshot
The Snapshot Custom Resource Definition (CRD) has a number of phases for the .status.phase field, including:
Snapshot Phase Description
New Not processed yet
Snapshotted Local snapshot was taken
SnapshotFailed Local snapshot was failed
Uploading The snapshot is being uploaded
Uploaded The snapshot is uploaded
UploadFailed The snapshot is failed to be uploaded
Canceling The upload of snapshot is being cancelled
Canceled The upload of snapshot is cancelled
CleanupAfterUploadFailed The Cleanup of local snapshot after the upload of snapshot was failed

Restore a Workload

Below is an example command of Velero restore. 
velero restore create --from-backup <velero-backup-name>

Velero restore will be marked as Completed when volume snapshots and other Kubernetes metadata have been successfully restored to the current cluster. At this point, all tasks of vSphere plugin related to this restore are completed as well. There are no any asynchronous data movement tasks behind the scene as that in the case of Velero backup.

CloneFromSnapshots

To restore from each volume snapshot, a CloneFromSnapshot Custom Resource (CR) will be created in the same namespace as the PVC that is originally snapshotted. We can get all CloneFromSnapshots in PVC namespace by running the following command. 
kubectl -n <pvc namespace> get clonefromsnapshot
CloneFromSnapshot CRD has some key phases for the .status.phase field:
Snapshot Phase Description
New Clone from snapshot is not completed
InProgress The vSphere volume snapshot is being downloaded from remote repository
Completed Clone from snapshot is completed
Failed Clone from snapshot is failed