Backing up the cluster

Prerequisites

Before taking a backup, you must provide the objectstore configuration to store the backup. For this, see Configuring the backup store.

Enabling the schedule snapshot backup

When configuring the schedule snapshot backup, it is recommended to provide the schedule and retention of the backup via the --schedule and --retention flags. By default, Automation Suite takes a backup every 45 minutes after starting and retains the last seven days of snapshots.

To enable the backup at the scheduled time, run the following command:

./uipathctl snapshot backup enable --schedule "*/45 * * * *" --retention 168h --prefix "mysnapshot"

Running this command ensure that backups are taken in scheduled time intervals.

Flag	Description
schedule	UNIX cron expression for schedule. This is only required if you want to enable a schedule backup. Default value is "*/45 * * * *", which means a backup is taken every 45 minutes.
retention	Retention policy in the following duration format: Hh:Mm:Ss. For example, 8h5m1s.
prefix	This will prefix the snapshot backups' names with the given string. This is only used if you want to enable a scheduled backup. The default value is a snapshot.

Important:

Note 1: Any cron expression provided as part of the --schedule parameter matches the time of the cluster node. We recommend syncing the cron expression with the scheduled backup of your external data stores (such as SQL database and objectstore).

Note 2: Automation Suite stores only the snapshot backed up during the defined retention policy. If your retention policy has a shorter duration, you may lose essential snapshots once the retention policy applies. Similarly, if the retention policy is longer, more snapshots are stored, which may take up space on your objectstore.

Note 3: Carefully consider the schedule setup. Taking snapshots at small intervals (for example, every 30 minutes) implies frequent backup operations, forcing you to store the last 30 minutes' data. Similarly, taking snapshots once a week can cause data loss if a disaster occurs much later than the previous backup. Therefore, we strongly recommend synchronizing the backup schedule and the retention duration with your Recovery Point Objective (RPO) requirements.

Disabling the schedule snapshot backup

To disable the scheduled backup, run the following command:

./uipathctl snapshot disable --prefix <prefix_name>

Note: You must provide the prefix you used while enabling the backup via the --prefix parameter; however, if you did not configure any prefix while allowing the backup, provide a snapshot as a prefix value.

On-demand snapshot backup

To take an on-demand snapshot backup, run the following command:

./uipathctl snapshot backup create <snapshot_name>

The aforementioned command takes an optional flag:

Flag	Description
--wait	Wait until the backup completes

If you do not provide the --wait command, the command executes in the background. To view the status of your backup, check the snapshot list as described in the following section.

Listing the existing snapshots

To list all the present snapshots, their status, and the time at which they were taken, run the following command:

./uipathctl snapshot list

Sample output:

NAME            STATUS           CREATION                       EXPIRATION
manualbackup1   Completed        2022-09-13 09:19:50 +0000 UTC  2023-09-13 09:19:50 +0000 UTC
manualbackup2   PartiallyFailed  2022-09-13 09:19:50 +0000 UTC  2023-09-13 09:19:50 +0000 UTC

Deleting a backup

Note:

Deleting a backup consists of deleting the cluster metadata and the volume data from the backup storage server. This operation is irreversible.

To delete the snapshot backups from the backup store, run the following command:

./uipathctl snapshot delete <snapshot_name>

You can retrieve the name of the snapshot that you want to delete by running the snapshot list command.

The command takes the following additional parameters:

Flag	Description
-a|--all	Delete all snapshots. If selecting this option, then you do not need to provide the name of the snapshot.
-f|--force	Forcibly delete the snapshot without confirmation.