Skip to main content

Understanding Backup and Restore

This topic provides an introduction to the Replicated KOTS snapshots feature for backup and restore. It describes how vendors enable snapshots, the type of data that is backed up, and how to troubleshoot issues for enterprise users.

note

The KOTS Snapshots feature is supported for existing cluster installations with KOTS and Replicated kURL installations only. Snapshots is not supported for Replicated Embedded Cluster installations. For more information about configuring backup and restore for Embedded Cluster, see Disaster Recovery for Embedded Cluster.

About Backup and Restore

You can enable Replicated KOTS snapshots to support backup and restore for existing cluster installations with KOTS and Replicated kURL installations.

Snapshots uses the Velero open source project as the backend to back up Kubernetes manifests and persistent volumes. Velero is a mature, fully-featured application. For more information, see the Velero documentation.

In addition to the default functionality that Velero provides, KOTS exposes hooks that let you inject scripts that can execute both before and after a backup, and before and after a restore. For more information, see Configuring Backup and Restore Hooks.

KOTS provides the Admin Console and the KOTS CLI so that your users can fully manage the backup and restore process. For more information, see Understanding Backup and Restore for Enterprise Users.

What Data is Backed Up

Full backups include the Admin Console and all application data, including KOTS-specific object-stored data. For Replicated kURL installations, this also backs up the Docker registry, which is required for air gapped installations.

Other Object-Stored Data

For embedded kURL clusters, you might be using object-stored data that is not specific to the kURL KOTS add-on.

For object-stored data that is not KOTS-specific and does not use persistentVolumeClaims (PVCs), you must write custom backup and restore hooks to enable back ups for that object-stored data. For example, Rook and Ceph do not use PVCs and so require custom backup and restore hooks. For more information about writing custom hooks, see Configuring Backup and Restore Hooks.

Pod Volume Data

Replicated supports only the restic backup program for pod volume data.

By default, Velero requires that you opt-in to have pod volumes backed up. In the Backup resource that you configure to enable snapshots, you must annotate each specific volume that you want to back up. For more information about including and excluding pod volumes, see Configuring Backups.

How to Enable Backup and Restore

To enable the snapshots backup and restore feature for your users, you must:

Understanding Backup and Restore for Enterprise Users

After vendors enable backup and restore, enterprise users install Velero and configure a storage destination in the Admin Console. Then users can create backups manually or schedule automatic backups. For more information about how users create and restore backups, see About Backing Up and Restoring with Snapshots.

Replicated recommends advising your users to make full backups for disaster recovery purposes. Additionally, full backups give users the flexibility to do a full restore, a partial restore (application only), or restore just the Admin Console.

From a full backup, users restore using the KOTS CLI or the Admin Console as indicated in the following table:

Restore TypeDescriptionInterface to Use
Full restoreRestores the Admin Console and the application.KOTS CLI
Partial restoreRestores the application only.KOTS CLI or Admin Console
Admin consoleRestores the Admin Console only.KOTS CLI

Partial backups are not recommended as they are a legacy feature and only back up the application volumes and manifests. Partial backups can be restored only from the Admin Console.

Troubleshooting Backup and Restore

To support end users with backup and restore, use the following resources: