pg_restorecluster

pg_restorecluster [options] version cluster backup

pg_restorecluster restores a PostgreSQL cluster from a backup created by pg_backupcluster. The cluster will be newly created in the system using the name provided on the command line; this allows renaming a cluster on restore. The restored cluster configuration will be updated to reflect the new name and location.

The backup name passed must end in either .basebackup or .dump; usually this will be the full path to a backup directory in /var/backups/postgresql/version-cluster/ as reported by pg_backupcluster ... list.

Basebackups are restored as-is. For dumps, pg_createcluster is used to create a new cluster, and schema and data are restored via pg_restore.

-d --datadir DIR

Use DIR as data directory for the restored cluster (default per createcluster.conf, by default /var/lib/postgresql/version/cluster).

-p --port N

Use port N for restored cluster (default is next free port).

-s --start

Start cluster after restoring (default for restore from dump; off for basebackup restores).

After the cluster has been started, ANALYZE is run on all databases.

--archive

Configure cluster for recovery from WAL archive. This sets restore_command to retrieve WAL files from backup/../wal.

--pitr TIMESTAMP

--recovery-target-time TIMESTAMP

Additionally to setting restore_command, set recovery_target_time to TIMESTAMP for point-in-time recovery. Also sets recovery_target_action='promote'.

--wal-archive DIR

For archive recovery, read WAL from archive DIR (default is backup/../wal).

/var/backups

Default root directory for cluster backup directories.

See pg_backupcluster(1) for a description of files.

pg_backupcluster(1), pg_restore(1), vacuumdb(1).

Christoph Berg <myon@debian.org>