Restore a MySQL Cluster from Backup

This guide explains how to restore a new MySQL cluster from an existing backup in KubeBlocks. It showcases two methods: using Cluster Annotation or the Ops API.

The Ops API provides enhanced control and progress monitoring capabilities for restore operations, acting as a wrapper around the Cluster Annotation.

Prerequisites

• KubeBlocks Environment:
  • KubeBlocks operator and required CRDs installed.
  • kubectl configured to access your Kubernetes cluster.
• Existing Backup:
  • A valid backup named example-mysql-backup-backup in the 'demo' namespace.

Verify Backup Status

Before restoring, ensure that a full backup is available. The restoration process will use this backup to create a new MySQL cluster.

Run the following command to check the backup status:

kubectl get backup -n demo

Expected Output:

NAME                           POLICY                                      METHOD       REPO      STATUS      TOTAL-SIZE   DURATION   DELETION-POLICY   CREATION-TIME          COMPLETION-TIME        EXPIRATION-TIME
example-mysql-cluster-backup   example-mysql-cluster-mysql-backup-policy   xtrabackup   s3-repo   Completed   1633717      18s        Delete            2025-03-07T03:25:22Z   2025-03-07T03:25:40Z

Restore the Backup Using Cluster Annotation

Create Restored Cluster

Create a new cluster with restoration configuration referencing the backup.

kubectl apply -f - <<EOF
apiVersion: apps.kubeblocks.io/v1
kind: Cluster
metadata:
  name: example-mysql-cluster-restored
  namespace: demo
  annotations:
    kubeblocks.io/restore-from-backup: '{"mysql":{"name":"example-mysql-cluster-backup","namespace":"demo","volumeRestorePolicy":"Parallel"}}'
spec:
  terminationPolicy: WipeOut
  componentSpecs:
    - name: mysql
      componentDef: "mysql-8.0"
      serviceVersion: 8.0.35
      disableExporter: false
      replicas: 2
      resources:
        limits:
          cpu: '0.5'
          memory: 0.5Gi
        requests:
          cpu: '0.5'
          memory: 0.5Gi
      volumeClaimTemplates:
        - name: data
          spec:
            storageClassName: ""
            accessModes:
              - ReadWriteOnce
            resources:
              requests:
                storage: 20Gi
EOF

Monitor Restoration Progress

Watch Pod Initialization:

kubectl get pods -n demo -w

Expected Workflow:

1. Data Preparation Pods:

restore-preparedata-XXXXX-<hash>   0/1     Init:0/1   0          6s
restore-preparedata-XXXXX-<hash>   1/1     Running    0          12s
restore-preparedata-XXXXX-<hash>   0/1     Completed 0          20s

These pods copy backup data to Persistent Volumes (PVCs).

2. MySQL Cluster Pods:

example-mysql-cluster-restored-mysql-0     0/4     Pending        0          0s
example-mysql-cluster-restored-mysql-0     4/4     Running        0          20s

Pods initialize with restored data and start MySQL services.

Perform Restoration via Ops API

Alternatively, use the Ops API to initiate the restoration process:

Step 1: Create a Restore Request

kubectl apply -f - <<EOF
apiVersion: operations.kubeblocks.io/v1alpha1
kind: OpsRequest
metadata:
  name: example-mysql-cluster-restore
  namespace: demo
spec:
  clusterName: example-mysql-cluster-restored
  force: false
  restore:
    backupName: example-mysql-cluster-backup
    backupNamespace: demo
  type: Restore
EOF

Step 2: Monitor Ops Progress

Monitor the progress of the restoration operation:

kubectl get ops example-mysql-cluster-restore -n demo -w

Expected Output:

NAME                            TYPE      CLUSTER                         STATUS    PROGRESS   AGE
example-mysql-cluster-restore   Restore   example-mysql-cluster-restored   Running   -/-        55s
example-mysql-cluster-restore   Restore   example-mysql-cluster-restored   Succeed   -/-        3m3s

Verify the Restored Cluster Status

After the restoration process is complete, verify the status of the newly restored cluster:

kubectl get cluster example-mysql-cluster-restored -n demo

Example Output:

NAME                    CLUSTER-DEFINITION   TERMINATION-POLICY   STATUS    AGE
example-mysql-cluster-restored                        WipeOut               Running   3m2s

Cleanup

To remove all created resources, delete the MySQL cluster along with its namespace:

kubectl delete cluster example-mysql-cluster -n demo
kubectl delete cluster example-mysql-cluster-restored -n demo
kubectl delete ns demo

Summary

This guide demonstrated how to restore a MySQL cluster from an existing backup using either the Cluster Annotation or Ops API. By following these steps, you can efficiently restore critical data to a new MySQL cluster in KubeBlocks.