Batch Backup and Restore Overview

Sometimes, an application may consist of multiple co-related components. For example, to deploy a WordPress, you will need a Deployment for the WordPress and another Deployment for the database. Now, it is sensible to want to backup or restore both of the deployments using a single configuration as they are parts of the same application.

Stash 0.9.0+ supports taking backup multiple co-related components using a single configuration known as BackupBatch. Stash 0.10.0+ supports restoring multiple co-related components together known as RestoreBatch This guide will give you an overview of how batch backup and restore works in Stash.

How Batch Backup Works

The following diagram shows how Stash takes backup of multiple co-related components in a single application. Open the image in a new tab to see the enlarged version.

Stash Batch Backup Flow
Fig: Batch backup flow in Stash

The backup process consists of the following steps:

  1. At first, a user creates a backend Secret. This secret holds the credentials to access the backend where the backed up data will be stored.

  2. Then, she creates a Repository crd which represents the original repository in the backend.

  3. Then, she creates a BackupBatch crd which specifies multiple targets(workload, volume, and database). It also specifies the Repository object that holds the backend information where the backed up data will be stored.

  4. Stash operator watches for BackupBatch objects.

  5. When it finds a BackupBatch object, it checks if there is any workload as a target. If there any, it injects a sidecar named stash into the workloads.

  6. It also creates a CronJob to trigger backups periodically.

  7. TheCronJob triggers backup on each scheduled slot by creating a BackupSession crd.

  8. The BackupSession controller (inside sidecar for sidecar model or inside the operator itself for job model) watches for BackupSession crd.

  9. When it finds a BackupSession it starts the backup process immediately(for job model a job is created for taking backup) for the individual targets. Stash operator enforces the backup order if the executionOrder is set to Sequential.

  10. The individual targets complete their backup process independently and update their respective fields in BackupSession status.

How Batch Restore Works

The following diagram shows the batch restore process. Please, open image in new tab to view the enlarged image.

Stash Batch Backup Flow
Fig: Batch restore flow in Stash

The batch restore process consists of the following steps:

  1. At first, the user creates a RestoreBatch CR specifying the targets and the respective Repository where the backed up data has been stored.
  2. The Stash operator watches for the RestoreBatch CR.
  3. When the Stash operator finds a RestoreBatch CR, it executes the global PreRestore hooks. If there is no global PreRestore hook, Stash will skip this step.
  4. Then, it injects an init-container into the target that follows the sidecar model and creates a restore job for the targets that follow the job model. Stash operator enforces the restore order in this step if the executionOrder is set to Sequential.
  5. The restore init-container/job first execute their local PreRestore hooks. Then, restore their data and finally execute their PostRestore hooks.
  6. Finally, Stash operator executes the global PostRestore hooks. If there is not global PostRestore hook configured for this RestoreBatch, Stash will skip this step.

Next Steps

  • See a step by step guide to backup application with multiple co-related components here.