4.5. Upgrading Remote Databases from PostgreSQL 10 to 12

Red Hat Virtualization 4.4 uses PostgreSQL 12, while RHV 4.3 uses PostgreSQL 10. If either of your databases (Manager or Data Warehouse) is installed on a separate machine, you must perform the following procedure on each remote database before upgrading the Manager.

Tip

This procedure uses the pg_dump and pg_restore commands. For a complete list of options for these commands, enter pg_dump --help or pg_restore --help on a machine with PostgreSQL installed.

  1. Back up all remote databases. Use the following command on each remote database machine: 

    # pg_dump -f backup-file database-name

    The default name for the engine database is engine, and the default name for the Data Warehouse database is ovirt-engine-history.

    For example, on the machine hosting the Data Warehouse database, enter a command similar to the following: 

    # pg_dump -f dwh-backup.dump ovirt-engine-history

  2. Copy the backup files to an external location such as a USB drive.

    Note

    This process requires installing Red Hat Enterprise Linux 8 on the Manager machine and all hosts, so make sure that you copy the backup file to a safe location.

  3. Install Red Hat Enterprise Linux 8 on each database server, with the Red Hat Virtualization subscription attached.

    For detailed installation instructions, see Performing a standard RHEL installation.

    The server must meet the minimum Manager requirements.

  4. Enable the correct repositories on each database server. For detailed instructions, see Enabling the Red Hat Virtualization Manager Repositories for Red Hat Virtualization 4.4.

    Updates to the Red Hat Virtualization Manager are released through the Content Delivery Network.

  5. Install the PostgreSQL server package: 

    # yum install postgresql postgresql-server postgresql-contrib

  6. Restore the backup on the data warehouse database machine and the engine database machine: 

    # pg_restore -d database-name backup-file

  7. On the Manager machine, run engine-setup to configure the Manager: 

    # engine-setup

You can now update the hosts.