NFS Full Restore#
The procedure herein details the steps that are required to perform a full cross-cluster Disaster Recovery restore of an MSR instance, leveraging Velero and MinIO to perform a complete end-to-end backup. This method is intended for scenarios wherein both the Kubernetes configuration and all registry data (blobs) must be migrated to a separate cluster, regardless of whether the underlying storage is shared.
Mirantis recommends using this approach only for small MSR instances, as it requires that you copy all registry data (image blobs) into and out of the MinIO backup storage, which is time-consuming and resource-intensive. Full backup and restore is recommended only for MSR instances that have a small total size of image and artifact data. For production environments with large registries, use the NFS Metadata Restore to a different cluster method.
Prerequisites#
- A freshly provisioned Kubernetes cluster that is ready for application deployment and free of any existing state or active resources that conflict with the deployment of the restored MSR stack.
- Verify that a compatible CSI driver is installed and configured in your Kubernetes cluster. For Snapshot backup, the CSI driver must support snapshot operations for your storage provider.