Backup Destinations

Backups can be made to the local file system or a remote destination. Off-site (non-local) backups are recommended to reduce the risk of the loss of information.

  • Display available backup destinations with backup list.

  • Add a new backup destination with backup add <location-name> <URI>.

    See Network URI specification.

Local backups are stored on a separate backup volume and the localbackup destination is pre-configured.

  • Display the list of localbackups with backup list localbackup, for example:

    $ backup list localbackup
       localbackup:
          URI: file:///backups
          Backups:
             2016-06-21 13:33
             2016-06-21 13:16
    

If the backup volume is too small, it can be increased - see: Backup Size Considerations. If the localbackup destination is removed or renamed, an ISO file upgrade will no longer function. Therefore, it is imperative that this destination is not removed.

Example:

backup add myserverbackup sftp://user:password@server/path

Azure example:

backup add myserverbackup sftp://<storage_account_name>.<sftp_local_user>[:password]@<storage_account_name>/path
  • <storage_acct_name> is set up on Azure. The name of the Storage Account.

  • <sftp_local_user> is set up on Azure if you enable SFTP and add a local user within the Storage Account settings.

Backups to sftp require ssh key-based authentication to be setup. Refer to SSH key management for further details. If you select ssh key-based authentication that was set up without a password, you will be prompted to input the password.

platform@VOSS-UN-6:~$ backup add remote sftp://[email protected]/sftp
No password found. Do you want to use sshkeys? [y/N]: y

If a common remote backup point is to be used by all nodes in the cluster, the backup destination needs to be added to each node. This can be automated by using cluster remote execution, for example:

cluster run all backup add myserverbackup sftp://user:password@server/path

The creation of scheduled backups of all nodes is done for failover reasons.

Note

In a multinode configuration, while only the highest priority secondary node backup contains data, in a failover scenario, the new primary node will contain the restored backup data.