Recover a Kubernetes environment after a disaster

Recover a Kubernetes environment after a failure by choosing a recovery approach based on whether you need to restore the original cluster or restore only applications on a new cluster.

Important

Restoring virtual machines alone does not restore a functional Kubernetes cluster because etcd state and control plane consistency are not recovered.

Recovery Options

Select one of the following recovery approaches:

  • Restore the same cluster using etcd (cluster disaster recovery): Recovers the exact cluster state, including identity, control plane configuration, and cluster-scoped resources.

  • Restore workloads on a new cluster: Resumes application operations without restoring cluster identity.

Restore the Same Cluster Using etcd (Cluster Disaster Recovery)

Restore the original cluster using an etcd snapshot to recover cluster identity, control plane configuration, and cluster-scoped resources. This approach is helpful only when cluster state recovery is required.

Important

This operation restores the cluster to a previous point in time. Existing changes after the snapshot are not preserved.

To restore the same cluster, follow these steps:

  1. Rebuild the cluster infrastructure and install a new cluster with the same topology and architecture, including nodes, networking, load balancers, DNS, and storage.

  2. Restore the etcd snapshot. For more information see, Restoring a Kubernetes etcd Snapshot to a File System.

  3. After the restore completes, copy etcd snapshot to the cluster to recover the cluster state.

    Kubernetes cluster restore resources
  4. Validate system namespaces.

  5. Restore application data and workloads such as PVCs, virtual machine disks, and application-level resources. For more information, see Restoring/Migrating Kubernetes Applications Out of Place.

Restoring Workloads on a New Cluster

Restore applications and data on a new cluster when cluster identity is not required.

Prerequisites

The application backup, including Kubernetes objects and persistent data must be available.

To restore workloads on a new cluster, follow these steps:

  1. Install a new cluster, which initializes a new etcd instance.

  2. Restore application data and workloads, including PVCs, virtual machine disks, and application-level resources. For more information, see Restoring/Migrating Kubernetes Applications Out of Place.

  3. Validate workloads.

×

Loading...