Skip to main content
Version: Next

Backup and Recovery Procedure for CloudnativePG Postgresql

Procedure for making backups​

By default, backups are performed for all CloudnativePG Postgresql databases and are performed daily and are stored in MinIO object storage for 7 days. This ensures that in the event that the data on the persistent volume of the CloudnativePG database is lost, the data can be recovered using the backup stored in MinIO. Changed in, for instance, the name of the bucket where backups are stored, the backup frequency or the retention duration can be made in the following sections in the production.yaml file.

Postgresql​

cloudnative_postgresql:
  ...
  cluster:
    backups:
      enabled: true
      endpointURL: http://my-other-object-store:9000
      provider: s3
      s3:
        region: eu-west-1
        bucket: "my-fancy-bucket-name"
        path: "/path-inside-bucket"
      scheduledBackups:
        - name: weekly-backup
          schedule: "0 0 0 * * 0"
      retentionPolicy: "100d"

TimescaleDB​

Changes to TimescaleDB backups are applied in the production.yaml sections of the respective JDBC-connector service. For instance for the Grafana database, one can set:

radar_jdbc_connector_grafana:
  ...
  timescaledb:
    cluster:
      backups:
        enabled: true
        endpointURL: http://my-other-object-store:9000
        provider: s3
        s3:
          region: eu-west-1
          bucket: "my-fancy-bucket-name"
          path: "/path-inside-bucket"
        scheduledBackups:
          - name: weekly-backup
            schedule: "0 0 0 * * 0"
        retentionPolicy: "100d"

Procedure for restoring backups​

To restore a backup, you need to add a recovery: section to the cluster: section of the respective database service in the production.yaml file like so:

Postgresql​

cloudnative_postgresql:
  ...
  cluster:
    mode: recovery
    recovery:
      method: object_store

TimescaleDB​

For instance for the Grafana database, one can set:

radar_jdbc_connector_grafana:
  ...
  timescaledb:
    cluster:
      mode: recovery
      recovery:
        method: object_store