Skip to content

Best practices for backup and restore

Review considerations and best practices that apply to application backup and restore.

Backup storage

  • Ensure that you have enough free space to receive and store backups. Running low on available disk space results in errors and affects system performance.
  • Regularly back up the production environment and potentially the system from the initial deployment.
  • Store backups on a machine other than the Control Center master. This can be a file server, a separate disaster recovery system, - or test environment.
  • To provide a historical archive, back up on a regular schedule. Back up as needed when you perform less-frequent tasks such as moving data from one instance to another or duplicating an instance for testing or failover purposes.

Restoring from a backup

  • Before upgrading or testing an application, ensure that you have a recent backup that successfully restores.
  • Frequently test restoring from a backup to ensure that the backup restores successfully, and that the restored system is an accurate representation of the state of the deployment when the backup was performed.
  • You can restore a backup to the system on which it was created or to an alternate system. When restoring a backup from one system to an alternate system, ensure that
    • The alternate system mirrors at least one device from the backed-up system.
    • Services that were added to the alternate system by a previous restore have been manually deleted.
  • Restoring from a backup file will remove services that were added after taking the backup. That is, if you create a backup, add a service, and then restore from the backup, the service is deleted as part of the restore process.
  • If an outage occurs during a restore from a backup, you can resume the restore because Control Center preserves complete data on the system. For example, if two of six backed up snapshots are restored before an outage, when you resume the restore, those two snapshots are saved on the system, and are not downloaded again.

Version considerations

  • Backups that were created using Control Center 1.0.x cannot be restored in Control Center 1.1.x or later.
  • In Control Center 1.2.x and later, you can restore backups that were created using Control Center 1.1.x or later.

Backup contents

Control Center backups contain:

  • Any applications installed deployed in Control Center
  • The Docker images used by that application
  • The contents of the Distributed File System (DFS)

Control Center backups do not contain:

  • The Control Center audit logs
  • The Control Center Elastic database where Resource Pools and their member delegates are defined
  • The Control Center configuration file, /etc/default/serviced
  • The serviced journal from the Control Center master host or any delegate hosts
  • The serviced .rpm or the directory structures created by it (i.e., serviced must be installed before a backup can be restored)