Skip to main content

Back up and restore cluster

note

Creating a backup strategy, testing that strategy, and performing backups is each customer's responsibility.

CelerData supports using Cluster Snapshot for disaster recovery on Elastic clusters.

The fundamental idea of disaster recovery for Elastic clusters is to ensure that the full cluster state (including data and metadata) is stored in object storage. This way, if the cluster encounters a failure, it can be restored from the object storage as long as the data and metadata remain intact. Additionally, features like backups and cross-region replication offered by cloud providers can be used to achieve remote recovery and cross-region disaster recovery.

In Elastic clusters, the data (Compute Node state) is stored in object storage, but the metadata (Coordinator Node state) remains local. To ensure that object storage has all the cluster state for restoration, CelerData supports Cluster Snapshot for both data and metadata in object storage.

note

Currently, only AWS-based and GCP-based Elastic clusters support backup and restoration with snapshots.

Enable Automated Cluster Snapshot​

Follow these steps to enable Automated Cluster Snapshot for your Elastic cluster:

note

You can only enable Automated Cluster Snapshot while the cluster is in the Running state.

  1. Sign in to the CelerData Cloud BYOC console.

  2. On the Clusters page, click the cluster that you want to enable Automated Cluster Snapshot for.

  3. On the cluster details page, click the Backup and Restore tab.

  4. Click Enable in the Automated Cluster Snapshot section.

  5. In the dialog box that appears, configure as follows, and click Confirm:

    a. Specify the Snapshot Interval. The default interval is 10 minutes.

    b. Select a storage volume for storing the snapshot files in the Storage volume drop-down list. The default storage volume is builtin_storage_volume, which is created during cluster deployment.

    note

    If you want to select a different storage volume other than builtin_storage_volume, make sure the credential that the storage volume referenced has the permissions to upload, access, and delete objects in the corresponding bucket. Meanwhile, make sure the data credential used for the cluster deployment has the same permissions on the corresponding bucket.

    The required permissions are listed below:

    • For AWS, the permissions are PutObject, GetObject, and DeleteObject.
    • For GCP, the permissions are storage.objects.create, storage.objects.get, and storage.objects.delete.

Restore cluster from snapshot​

Follow these steps to restore your Elastic cluster from the snapshot:

note

You can only restore the cluster from the snapshot while the cluster is in the Suspended state.

  1. Sign in to the CelerData Cloud BYOC console.
  2. On the Clusters page, click the cluster that you want to restore.
  3. On the cluster details page, click the Backup and Restore tab.
  4. Choose the snapshot from which you want restore the cluster in the Automated Cluster Snapshot section, and click the Restore button.
  5. On the dialog box that appears, click Confirm.

Restoring a cluster will take a few minutes. The cluster will enter the Updating state for recovery, and then the Running state when the recovery is complete.