Recovery

In PostgreSQL terminology, recovery is the process of starting a PostgreSQL instance using a previously taken backup. The PostgreSQL recovery mechanism is very solid and rich. It also supports Point In Time Recovery, which allows you to restore a given cluster up to any point in time from the first available backup in your catalog to the last archived WAL (as you can see, the WAL archive is mandatory in this case).

In CloudNativePG, recovery cannot be performed "in-place" on an existing cluster. Recovery is rather a way to bootstrap a new Postgres cluster starting from an available physical backup.

Note

For details on the bootstrap stanza, please refer to the "Bootstrap" section.

The recovery bootstrap mode lets you create a new cluster from an existing physical base backup, and then reapply the WAL files containing the REDO log from the archive.

WAL files are pulled from the defined recovery object store.

Base backups may be taken either on object stores, or using volume snapshots (from version 1.21).

Warning

Recovery using volume snapshots had an initial release on 1.20.1. Because of the amount of progress on the feature for 1.21.0, it is strongly advised that you upgrade to 1.21.0 or more advanced releases to use volume snapshots.

Recovery from a recovery object store can be achieved in two ways:

  • using a recovery object store, that is, a backup of another cluster created by Barman Cloud and defined via the barmanObjectStore option in the externalClusters section (recommended)
  • using an existing Backup object in the same namespace (this was the only option available before version 1.8.0).

Both recovery methods enable either full recovery (up to the last available WAL) or up to a point in time. When performing a full recovery, the cluster can also be started in replica mode (see replica clusters for reference). If using replica mode, make sure that the PostgreSQL configuration (.spec.postgresql.parameters) of the recovered cluster is compatible, from a physical replication standpoint, with the original one.

For recovery using volume snapshots:

  • using a consistent set of VolumeSnapshot objects that all belong to the same backup, and identified by the same cnpg.io/cluster and cnpg.io/backupName labels, then recovering through the volumeSnapshots option in the .spec.bootstrap.recovery stanza, as described in "Recovery from VolumeSnapshot objects" below

Recovery from an object store

You can recover from a backup created by Barman Cloud and stored on a supported object store. Once you have defined the external cluster, including all the required configuration in the barmanObjectStore section, you need to reference it in the .spec.recovery.source option. The following example defines a recovery object store in a blob container in Azure:

apiVersion: postgresql.cnpg.io/v1
kind: Cluster
metadata:
  name: cluster-restore
spec:
  [...]

  superuserSecret:
    name: superuser-secret

  bootstrap:
    recovery:
      source: clusterBackup

  externalClusters:
    - name: clusterBackup
      barmanObjectStore:
        destinationPath: https://STORAGEACCOUNTNAME.blob.core.windows.net/CONTAINERNAME/
        azureCredentials:
          storageAccount:
            name: recovery-object-store-secret
            key: storage_account_name
          storageKey:
            name: recovery-object-store-secret
            key: storage_account_key
        wal:
          maxParallel: 8

Important

By default the recovery method strictly uses the name of the cluster in the externalClusters section as the name of the main folder of the backup data within the object store, which is normally reserved for the name of the server. You can specify a different folder name with the barmanObjectStore.serverName property.

Note

In the above example we are taking advantage of the parallel WAL restore feature, dedicating up to 8 jobs to concurrently fetch the required WAL files from the archive. This feature can appreciably reduce the recovery time. Make sure that you plan ahead for this scenario and correctly tune the value of this parameter for your environment. It will certainly make a difference when (not if) you'll need it.

Recovery from VolumeSnapshot objects

Warning

When creating replicas after having recovered the primary instance from the volume snapshot, the operator might end up using pg_basebackup to synchronize them, resulting in a slower process depending on the size of the database. This limitation will be lifted in the future when support for online backups and PVC cloning will be introduced.

CloudNativePG can create a new cluster from a VolumeSnapshot of a PVC of an existing Cluster that's been taken using the declarative API for volume snapshot backups. You will need to specify the name of the snapshot, as in the following example:

apiVersion: postgresql.cnpg.io/v1
kind: Cluster
metadata:
  name: cluster-restore
spec:
  [...]

  bootstrap:
    recovery:
      volumeSnapshots:
        storage:
          name: <snapshot name>
          kind: VolumeSnapshot
          apiGroup: snapshot.storage.k8s.io

In case the backed-up cluster was using a separate PVC to store the WAL files, the recovery must include that too:

apiVersion: postgresql.cnpg.io/v1
kind: Cluster
metadata:
  name: cluster-restore
spec:
  [...]

  bootstrap:
    recovery:
      volumeSnapshots:
        storage:
          name: <snapshot name>
          kind: VolumeSnapshot
          apiGroup: snapshot.storage.k8s.io

        walStorage:
          name: <snapshot name>
          kind: VolumeSnapshot
          apiGroup: snapshot.storage.k8s.io

Warning

If bootstrapping a replica-mode cluster from snapshots, to leverage snapshots for the standby instances and not just the primary, it would be advisable to:

  1. start with a single instance replica cluster. The primary instance will be recovered using the snapshot and available WALs form the source cluster
  2. take a snapshot of the primary in the replica cluster
  3. increase the number of instances in the replica cluster as desired

Recovery from a Backup object

In case a Backup resource is already available in the namespace in which the cluster should be created, you can specify its name through .spec.bootstrap.recovery.backup.name, as in the following example:

apiVersion: postgresql.cnpg.io/v1
kind: Cluster
metadata:
  name: cluster-example-initdb
spec:
  instances: 3

  bootstrap:
    recovery:
      backup:
        name: backup-example

  storage:
    size: 1Gi

This bootstrap method allows you to specify just a reference to the backup that needs to be restored.

The previous example implies the application database and its owning user to be the default one, app. If the PostgreSQL cluster being restored was using different names, they can be specified as documented in the Configure the application database section.

Additional considerations

Whether you recover from a recovery object store, a volume snapshot, or an existing Backup resource, the following considerations apply:

  • The application database name and the application database user are preserved from the backup that is being restored. The operator does not currently attempt to back up the underlying secrets, as this is part of the usual maintenance activity of the Kubernetes cluster itself.
  • To preserve the original postgres user password, you need to properly configure enableSuperuserAccess and supply a superuserSecret.
  • By default, the recovery will continue up to the latest available WAL on the default target timeline (current for PostgreSQL up to 11, latest for version 12 and above). You can optionally specify a recoveryTarget to perform a point in time recovery (see the "Point in time recovery" section).

Important

Consider using the barmanObjectStore.wal.maxParallel option to speed up WAL fetching from the archive by concurrently downloading the transaction logs from the recovery object store.

Point in time recovery (PITR)

Instead of replaying all the WALs up to the latest one, we can ask PostgreSQL to stop replaying WALs at any given point in time, after having extracted a base backup. PostgreSQL uses this technique to achieve point-in-time recovery (PITR). The presence of a WAL archive is mandatory.

Important

PITR requires you to specify a recovery target, by using the options described in the "Recovery targets" section below.

The operator will generate the configuration parameters required for this feature to work in case a recovery target is specified.

PITR from an object store

The example below uses a recovery object store in Azure that contains both the base backups and the WAL archive. The recovery target is based on a requested timestamp:

apiVersion: postgresql.cnpg.io/v1
kind: Cluster
metadata:
  name: cluster-restore-pitr
spec:
  instances: 3

  storage:
    size: 5Gi

  bootstrap:
    recovery:
      # Recovery object store containing WAL archive and base backups
      source: clusterBackup
      recoveryTarget:
        # Time base target for the recovery
        targetTime: "2023-08-11 11:14:21.00000+02"

  externalClusters:
    - name: clusterBackup
      barmanObjectStore:
        destinationPath: https://STORAGEACCOUNTNAME.blob.core.windows.net/CONTAINERNAME/
        azureCredentials:
          storageAccount:
            name: recovery-object-store-secret
            key: storage_account_name
          storageKey:
            name: recovery-object-store-secret
            key: storage_account_key
        wal:
          maxParallel: 8

You might have noticed that in the above example you only had to specify the targetTime in the form of a timestamp, without having to worry about specifying the base backup from which to start the recovery.

The backupID option is the one that allows you to specify the base backup from which to initiate the recovery process. By default, this value is empty.

If you assign a value to it (in the form of a Barman backup ID), the operator will use that backup as base for the recovery.

Important

You need to make sure that such a backup exists and is accessible.

If the backup ID is not specified, the operator will automatically detect the base backup for the recovery as follows:

  • when you use targetTime or targetLSN, the operator selects the closest backup that was completed before that target
  • otherwise the operator selects the last available backup in chronological order.

PITR from VolumeSnapshot Objects

The example below uses:

  • a Kubernetes volume snapshot for the PGDATA containing the base backup from which to start the recovery process, identified in the recovery.volumeSnapshots section and called test-snapshot-1
  • a recovery object store in MinIO containing the WAL archive, identified by the recovery.source option in the form of an external cluster definition

The recovery target is based on a requested timestamp.

apiVersion: postgresql.cnpg.io/v1
kind: Cluster
metadata:
  name: cluster-example-snapshot
spec:
  # ...
  bootstrap:
    recovery:
      source: cluster-example-with-backup
      volumeSnapshots:
        storage:
          name: test-snapshot-1
          kind: VolumeSnapshot
          apiGroup: snapshot.storage.k8s.io
      recoveryTarget:
        targetTime: "2023-07-06T08:00:39"
  externalClusters:
    - name: cluster-example-with-backup
      barmanObjectStore:
        destinationPath: s3://backups/
        endpointURL: http://minio:9000
        s3Credentials:
          accessKeyId:
            name: minio
            key: ACCESS_KEY_ID
          secretAccessKey:
            name: minio
            key: ACCESS_SECRET_KEY

Note

In case the backed up Cluster had walStorage enabled, you also must specify the volume snapshot containing the PGWAL directory, as mentioned in the Recovery from VolumeSnapshot objects section.

Warning

It is your responsibility to ensure that the end time of the base backup in the volume snapshot is prior to the recovery target timestamp.

Recovery targets

Here are the recovery target criteria you can use:

targetTime
time stamp up to which recovery will proceed, expressed in RFC 3339 format (the precise stopping point is also influenced by the exclusive option)
targetXID
transaction ID up to which recovery will proceed (the precise stopping point is also influenced by the exclusive option); keep in mind that while transaction IDs are assigned sequentially at transaction start, transactions can complete in a different numeric order. The transactions that will be recovered are those that committed before (and optionally including) the specified one
targetName
named restore point (created with pg_create_restore_point()) to which recovery will proceed
targetLSN
LSN of the write-ahead log location up to which recovery will proceed (the precise stopping point is also influenced by the exclusive option)
targetImmediate
recovery should end as soon as a consistent state is reached - i.e. as early as possible. When restoring from an online backup, this means the point where taking the backup ended

Important

While the operator is able to automatically retrieve the closest backup when either targetTime or targetLSN is specified, this is not possible for the remaining targets: targetName, targetXID, and targetImmediate. In such cases, it is important to specify backupID, unless you are OK with the last available backup in the catalog.

The example below uses a targetName based recovery target:

apiVersion: postgresql.cnpg.io/v1
kind: Cluster
[...]
  bootstrap:
    recovery:
      source: clusterBackup
      recoveryTarget:
        backupID: 20220616T142236
        targetName: 'restore_point_1'
[...]

You can choose only a single one among the targets above in each recoveryTarget configuration.

Additionally, you can specify targetTLI force recovery to a specific timeline.

By default, the previous parameters are considered to be inclusive, stopping just after the recovery target, matching the behavior in PostgreSQL You can request exclusive behavior, stopping right before the recovery target, by setting the exclusive parameter to true like in the following example relying on a blob container in Azure for both base backups and the WAL archive:

apiVersion: postgresql.cnpg.io/v1
kind: Cluster
metadata:
  name: cluster-restore-pitr
spec:
  instances: 3

  storage:
    size: 5Gi

  bootstrap:
    recovery:
      source: clusterBackup
      recoveryTarget:
        backupID: 20220616T142236
        targetName: "maintenance-activity"
        exclusive: true

  externalClusters:
    - name: clusterBackup
      barmanObjectStore:
        destinationPath: https://STORAGEACCOUNTNAME.blob.core.windows.net/CONTAINERNAME/
        azureCredentials:
          storageAccount:
            name: recovery-object-store-secret
            key: storage_account_name
          storageKey:
            name: recovery-object-store-secret
            key: storage_account_key
        wal:
          maxParallel: 8

Configure the application database

For the recovered cluster, we can configure the application database name and credentials with additional configuration. To update application database credentials, we can generate our own passwords, store them as secrets, and update the database use the secrets. Or we can also let the operator generate a secret with randomly secure password for use. Please reference the "Bootstrap an empty cluster" section for more information about secrets.

The following example configure the application database app with owner app, and supplied secret app-secret.

apiVersion: postgresql.cnpg.io/v1
kind: Cluster
[...]
spec:
  bootstrap:
    recovery:
      database: app
      owner: app
      secret:
        name: app-secret
      [...]

With the above configuration, the following will happen after recovery is completed:

  1. if database app does not exist, a new database app will be created.
  2. if user app does not exist, a new user app will be created.
  3. if user app is not the owner of database, user app will be granted as owner of database app.
  4. If value of username match value of owner in secret, the password of application database will be changed to the value of password in secret.

Important

For a replica cluster with replica mode enabled, the operator will not create any database or user in the PostgreSQL instance, as these will be recovered from the original cluster.

How recovery works under the hood

You can use the data uploaded to the object storage to bootstrap a new cluster from a previously taken backup. The operator will orchestrate the recovery process using the barman-cloud-restore tool (for the base backup) and the barman-cloud-wal-restore tool (for WAL files, including parallel support, if requested).

For details and instructions on the recovery bootstrap method, please refer to the "Bootstrap from a backup" section.

Important

If you are not familiar with how PostgreSQL PITR works, we suggest that you configure the recovery cluster as the original one when it comes to .spec.postgresql.parameters. Once the new cluster is restored, you can then change the settings as desired.

Under the hood, the operator will inject an init container in the first instance of the new cluster, and the init container will start recovering the backup from the object storage.

Important

The duration of the base backup copy in the new PVC depends on the size of the backup, as well as the speed of both the network and the storage.

When the base backup recovery process is completed, the operator starts the Postgres instance in recovery mode: in this phase, PostgreSQL is up, albeit not able to accept connections, and the pod is healthy according to the liveness probe. Through the restore_command, PostgreSQL starts fetching WAL files from the archive (you can speed up this phase by setting the maxParallel option and enable the parallel WAL restore capability).

This phase terminates when PostgreSQL reaches the target (either the end of the WAL or the required target in case of Point-In-Time-Recovery). Indeed, you can optionally specify a recoveryTarget to perform a point in time recovery. If left unspecified, the recovery will continue up to the latest available WAL on the default target timeline (current for PostgreSQL up to 11, latest for version 12 and above).

Once the recovery is complete, the operator will set the required superuser password into the instance. The new primary instance will start as usual, and the remaining instances will join the cluster as replicas.

The process is transparent for the user and it is managed by the instance manager running in the Pods.

Restoring into a cluster with a backup section

A manifest for a cluster restore may include a backup section. This means that the new cluster, after recovery, will start archiving WAL's and taking backups if configured to do so.

For example, the section below could be part of a manifest for a Cluster bootstrapping from Cluster cluster-example-backup, and would create a new folder in the storage bucket named recoveredCluster where the base backups and WAL's of the recovered cluster would be stored.

  backup:
    barmanObjectStore:
      destinationPath: s3://backups/
      endpointURL: http://minio:9000
      serverName: "recoveredCluster"
      s3Credentials:
        accessKeyId:
          name: minio
          key: ACCESS_KEY_ID
        secretAccessKey:
          name: minio
          key: ACCESS_SECRET_KEY
    retentionPolicy: "30d"

  externalClusters:
  - name: cluster-example-backup
    barmanObjectStore:
      destinationPath: s3://backups/
      endpointURL: http://minio:9000
      s3Credentials:

You should not re-use the exact same barmanObjectStore configuration for different clusters. There could be cases where the existing information in the storage buckets could be overwritten by the new cluster.

Warning

The operator includes a safety check to ensure a cluster will not overwrite a storage bucket that contained information. A cluster that would overwrite existing storage will remain in state Setting up primary with Pods in an Error state. The pod logs will show: ERROR: WAL archive check failed for server recoveredCluster: Expected empty archive

Important

If you set the cnpg.io/skipEmptyWalArchiveCheck annotation to enabled in the recovered cluster, you can skip the above check. This is not recommended as for the general use case the above check works fine. Please don't do this unless you are familiar with PostgreSQL recovery system, as this can lead you to severe data loss.