Configure backup storage

Before configuring backups, you must set up a backup storage location.

1. Create a Kubernetes secret using your cloud storage credentials.

   apiVersion: v1
   data:
       AWS_ACCESS_KEY_ID: YOUR_ACCESS_KEY_ID_BASE64_ENCODED
       AWS_SECRET_ACCESS_KEY: YOUR_SECRET_ACCESS_KEY_BASE64_ENCODED
   kind: Secret
   metadata:
       name: my-s3-backup-storage
   type: Opaque
   

2. Create a BackupStorage CRD that references this secret.

   apiVersion: everest.percona.com/v1alpha1
   kind: BackupStorage
   metadata:
       name: my-s3-backup-storage
   spec:
       bucket: my-s3-bucket
       credentialsSecretName: my-s3-backup-storage
       description: My S3 backup storage
       endpointURL: https://my-s3-endpoint.com
       forcePathStyle: false
       region: us-west-2
       type: s3
       verifyTLS: true
   

3. Configure backup schedules in your DatabaseCluster.

   spec:
     backup:
       schedules:
         - name: "daily-backup"
           enabled: true
           schedule: "0 0 * * *"  # Daily at midnight
           retentionCopies: 7
           backupStorageName: "my-s3-backup-storage"
       pitr:  # Point-in-Time Recovery
         enabled: true
         backupStorageName: "my-s3-backup-storage"
         uploadIntervalSec: 300  # 5 minutes
   

Manual backups and restores

In addition to scheduled backups, you can create manual backups and perform restores using the DatabaseClusterBackup and DatabaseClusterRestore CRDs.

Create a manual backup

Define a DatabaseClusterBackup object:

apiVersion: everest.percona.com/v1alpha1
kind: DatabaseClusterBackup
metadata:
  labels:
    clusterName: my-database-cluster
  name: my-database-cluster-backup
spec:
  backupStorageName: my-s3-backup-storage
  dbClusterName: my-database-cluster

Monitor the backup status:

kubectl get dbbackup manual-backup-2024-04-11 -o jsonpath='{.status}'

A backup typically progresses through these states:

• Starting: Initial backup preparation

• Running: Backup in progress

• Succeeded: Backup completed successfully

• Failed: Backup failed

Restore from a backup

You can restore from a backup in two ways:

1. Restore from a DatabaseClusterBackup.

   apiVersion: everest.percona.com/v1alpha1
   kind: DatabaseClusterRestore
   metadata:
     name: restore-from-backup
   spec:
     dbClusterName: my-database-cluster
     dataSource:
       dbClusterBackupName: my-database-cluster-backup
   

2. Restore to a new database cluster.

   To restore a backup to a new database cluster, create a new DatabaseCluster with the dataSource field that references the backup:

   apiVersion: everest.percona.com/v1alpha1
   kind: DatabaseCluster
   metadata:
     name: restored-database
     namespace: default
   spec:
     # .. hidden
     dataSource:
       dbClusterBackupName: backup-name-here  # Name of the backup to restore from
   

Important

• The new cluster must use the same type and version of database engine as the backup.
• The storage size of the new cluster should be at least as large as that of the original cluster.
• The new cluster will be created in the same namespace as the backup.
• Other configurations, such as proxy settings and monitoring, may differ from those of the original cluster.

Perform Point-in-Time Recovery (PITR)

You can perform a point-in-time recovery to restore your database to a specific moment in time.

apiVersion: everest.percona.com/v1alpha1
kind: DatabaseClusterRestore
metadata:
  name: pitr-restore
spec:
  dbClusterName: my-database
  dataSource:
    dbClusterBackupName: base-backup
    pitr:
      type: date
      date: "2024-04-11T15:30:00Z"  # UTC timestamp

Monitor the restore status:

kubectl get dbrestore restore-from-backup -o jsonpath='{.status}'

The restore will typically go through these states:

• Starting: Initial restore preparation
• Restoring: Restore in progress
• Succeeded: Restore completed successfully
• Failed: Restore failed