PostgreSQL

maintenance-mode-start

Put the cluster into maintenance mode, where Juju hooks will do nothing and the charm will make no changes to the deployment. This action must be run on the Juju leader. It is your responsibility to ensure that any manual maintenance work you perform will not confuse the charm and it will regain control once maintenance mode is ended. Turning off mainenance mode after significant changes, such as a manual failover, can result in catastrophic dataloss. It is also important to not instruct Juju to make changes during maintenance mode, such as adding storage to units or upgrading the charm, as the charm will not react to these changes and the operations can also result in catestrophic dataloss when required preamble steps are not performed.

maintenance-mode-stop

See maintenance-mode-start.

replication-pause

Pause replication replay on a hot standby unit.

replication-resume

Resume replication replay on a hot standby unit.

switchover

Promote a specific unit to master. Must be run on the leader unit.

Params

• master string

  Unit to promote to master (eg. postgresql/3).

Required

master

wal-e-backup

Run a wal-e backup to cloud storage now. Requires WAL shipping to be enabled with wal-e. Action terminates when the backup is complete.

Params

• prune boolean

  Run the configured prune step if the backup completes successfully, the same as the regularly scheduled backup task.

wal-e-list-backups

List backups available for PITR and their metadata.

Params

• storage-uri string

  The WAL-E storage URI to search for backups.

wal-e-restore

PITR database recovery from configured wal-e store. THIS WILL DESTROY YOUR EXISTING DATA. Most of these options correspond to PostgreSQL recovery target settings, documented at http://www.postgresql.org/docs/current/static/recovery-target-settings.html.

Params

• backup-name string

  The name of the backup to recover from. Use the wal-e-list-backups action to see what is available. By default, the most recent backup is used. The selected backup must have been created before any specified target-time.

• confirm boolean

  Recovery destroys the current local database and PITR backups. If this option is false, the action will just report details of what will be destroyed. Set this option to true to perform the recovery.

• storage-uri string

  The WAL-E storage URI to recover from. If you are rewinding a deployment to an earlier point in time, this may be the same as the wal_e_storage_uri configuration option. See the WAL-E or wal_e_storage_uri configuration option documentation for supported syntax. Requires the relevant os_*, aws_* or wabs_* configuration options to be specified.

• target-time string

  Target time to recovery database to in ISO8601 format. By default recovers to the latest available.

• target-timeline string

  The timeline to recover to. Set to an empty string to recover along the same timeline that was current when the backup was taken. The default value 'latest' will to recover to the latest available timeline, following any promotions. Other values will recover along that specific timeline.

Required

storage-uri