XCO System Backup and Restoration

Manual Backup and Restore

XCO supports backup and restore across the same IP modes. A warning message appears during restore operation.

The following table describes the support matrix for backup and restore operation across the IP modes:

A backup generated on one XCO system can be restored on another system.

For more information, see Back up and Restore the XCO System.

You can use the efa system backup-list command to see the backup files that are available to use in a restore operation. For example:

$ efa system backup-list
+----+-----------------------------------+---------+----------------+
| ID    | File                           | Version | Generated By   |
+----+-----------------------------------+---------+----------------+
| 1     | EFA-2020.08.20-20.26.46.tar    | 2.3.0-1 | User           |
+----+-----------------------------------+---------+----------------+
| 2     | EFA-2020.08.20-20.27.29.tar    | 2.3.0-GA| System         |
+----+-----------------------------------+---------+----------------+
--- Time Elapsed: 183.69386ms ---

Periodic Backups and Configuration

XCO periodically backs up the system (12 AM by default, but configurable). During the backup process, all the services are locked and APIs return the message "Service is Locked with reason backup". Therefore, you cannot see the location under dropdown for Device Discovery.

The periodic backup process creates a backup tar file that is saved to the same location as the manual backup files. When new backup and supportsave files are created, according to the age of the backup files and maximum number of backup files to save, the system deletes the saved system-generated backup files, supportsave files, and manual backup files.

You can use the efa system settings show command to view the current backup settings.

Tip

You can use the efa system cleanup REST API to delete a specified backup or supportsave file. This feature lets you delete files before they are automatically deleted.

You can use the efa system settings reset command to reset the backup system settings to default values.

Backups During Upgrades

The upgrade process of XCO also backs up the XCO system to recover data if the upgrade fails. For more information, see "Recover from an Upgrade Failure" in the ExtremeCloud Orchestrator Deployment Guide, 3.6.0 .

Logs

The REST APIs for backup, restore, and supportsave are part of the Monitoring Service and can be accessed via port 8078. The logs for the REST APIs are saved to the <log_dir>/monitor/ location.

The efa-monitor service is available only on multi-node deployments. The logs contain:

• Galera and k3s recovery logs
• RabbitMQ recovery logs

The efa-monitor service checks status of clusters (galera, k3s or rabbitmq) every minute. If it detects any inconsistencies in the cluster, it will recover. The logs help in understanding the state of the cluster.

Use the sudo systemctl status/stop/start efamonitor to manage the monitoring service.

Before you manually stop any service for debugging purposes, ensure to stop the efa-monitor service. Otherwise, the efa-monitor service will automatically attempt to recover.

Enable or Disable Database Log Traces

You can use the efa fabric debug set command to set the debug mode for debugging purposes for the fabric service.

For information about commands and supported parameters to configure database log traces, see ExtremeCloud Orchestrator Command Reference, 3.6.0 .

The following table describes the database log trace settings in XCO 3.5.0 and XCO 3.6.0:

The following table describes the log level and the CLI output information: