Question: How do you restore a PostgreSQL cluster?


Restoring a PostgreSQL cluster involves several steps, depending on the backup method used. Here are two common scenarios: using pg_dump and pg_restore, and using physical backups.

Using pg_dump and pg_restore

  1. Backup with pg_dump: Back up your database cluster using pg_dump or pg_dumpall. For a full cluster backup with pg_dumpall, you might use:

    pg_dumpall > backup.sql
  2. Restore with pg_restore (if individual databases were backed up) or SQL execution (if pg_dumpall was used):

    • If the backup was created with pg_dumpall, restoring is straightforward:
      psql -f backup.sql postgres
    • If individual databases were backed up with pg_dump, use pg_restore:
      pg_restore -d dbname backup_file

Using Physical Backups

For large databases, physical backups can be more efficient. These involve copying the data directory of the PostgreSQL server.

  1. Create a Physical Backup: Stop your PostgreSQL server and copy the entire data directory to a backup location.

    systemctl stop postgresql cp -R /var/lib/postgresql/12/main /path/to/backup/ systemctl start postgresql
  2. Restore the Physical Backup:

    • Ensure PostgreSQL is stopped on the server where you are restoring the data.
    • Remove the current data directory and replace it with the backup:
      systemctl stop postgresql rm -rf /var/lib/postgresql/12/main cp -R /path/to/backup /var/lib/postgresql/12/main chown -R postgres:postgres /var/lib/postgresql/12/main systemctl start postgresql


  • Always ensure that the PostgreSQL versions between the backup source and destination match to avoid compatibility issues.
  • Test your backup and restoration process regularly to confirm everything works as expected.
  • Consider using tools like Barman or pgBackRest for more sophisticated backup and recovery operations, especially in production environments.

Was this content helpful?

White Paper

Free System Design on AWS E-Book

Download this early release of O'Reilly's latest cloud infrastructure e-book: System Design on AWS.

Free System Design on AWS E-Book
Start building today

Dragonfly is fully compatible with the Redis ecosystem and requires no code changes to implement.