Skip to main content
Version: 4.x

Data Recovery

important

Creating a data protection strategy, testing that strategy, and performing backups are each customer's responsibility.

PhoenixAI supports using Cluster Snapshot for data recovery on Elastic clusters.

The fundamental idea of data recovery for PhoenixAI 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.

In PhoenixAI 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, PhoenixAI supports Cluster Snapshot for storing both data and metadata in object storage.

note
  • Currently, this feature is available only on clusters deployed on AWS and GCP.
  • Currently, this feature is available only on clusters running on database kernel v4.0.10, v4.1.2, or later.

Configure recovery strategy​

Currently, PhoenixAI supports local recovery, that is, recovery in the original cluster or another cluster in the same cloud region.

important

You can only configure the recovery strategy while the cluster is in the Running state.

Follow these steps to configure the local recovery strategy for your PhoenixAI cluster:

  1. Sign in to the PhoenixAI Cloud console.
  2. On the Clusters page, click the cluster for which you want to configure the recovery strategy.
  3. On the Data Protection tab of the cluster details page, click Configure in the Local Recovery section.
  4. On the Configure Local Recovery page, select a Snapshot schedule for the cluster snapshot, and click Enable Local Recovery.

After the recovery strategy is enabled, the system starts to generate the first cluster snapshot. Generating a cluster snapshots will take a few minutes, depending on the cluster size.

note

The cluster snapshot files are stored in the bucket referenced in the cluster's Data Credential, and they are stored to your object storage via builtin_storage_volume, that is, the storage volume created during the cluster deployment.

Restore cluster from snapshot​

Currently, you can restore the data and metadata to the current cluster or another cluster in the same cloud region.

important

You can only restore a cluster while the cluster is in the Running state.

Restore to current cluster​

Destructive

Restoring data to the current cluster will overwritten all existing data with the data in the snapshot.

Follow these steps to restore your data and metadata to the current cluster:

  1. Sign in to the PhoenixAI Cloud console.

  2. On the Clusters page, click the cluster that you want to restore.

  3. On the Data Protection tab of the cluster details page, click Restore from Snapshot in the Local Recovery section.

  4. On the Restore from Snapshot page, configure as follows, and click Next:

    Step 1: Select Snapshot

    Select the snapshot file for the restore operation.

    Step 2: Choose Restore Target

    Select Restore to Current Cluster.

  5. On the Confirm Restore to Current Cluster page, review the information, enter the cluster name to confirm the operation, select I understand that all data written after the snapshot will be permanently lost and I understand the the cluster will be unavailable during the restore, and click Restore Cluster.

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.

Restore to another cluster in the same cloud region​

Non-destructive

Restoring data to another cluster will not affect the existing data in the original cluster.

  1. Sign in to the PhoenixAI Cloud console.

  2. On the Clusters page, click the cluster that you want to restore.

  3. On the Data Protection tab of the cluster details page, click Restore from Snapshot in the Local Recovery section.

  4. On the Restore from Snapshot page, configure as follows, and click Next:

    Step 1: Select Snapshot

    Select the snapshot file for the restore operation.

    Step 2: Choose Restore Target

    Select Restore to Existing Cluster, and then select an existing cluster as the target cluster or Create a new cluster.

    If you select Create a new cluster, you will be first redirected to the PhoenixAI cluster Deployment Wizard page. After the new cluster is successfully deployed, click Back to restore in the STEP4 of the Deployment Wizard. You will then be redirect to the Data Protection tab.

    important

    The target cluster must be created with the same Data Credential and Deployment Credential as the original cluster.

  5. On the Confirm Restore to Existing Cluster page, review the information, select I understand that existing data on cluster will be replaced with the snapshot data, and click Restore to Cluster.

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