NetVault 13.0.1 - Administration Guide for Managed Service Providers

The procedure for performing backups using the Plug-in for FileSystem is similar for both standard and virtual clients. The cluster-aware version of the plug-in can be used in the same manner as the standard version to select the data items. However, when you open the virtual client node on the NetVault Selections page, only the cluster-aware plug-in is listed under the node. The plug-in displays both shared and local drives and mount points in the selection tree. You must make a note of the drive letter or mount point for the shared resource and select the data accordingly. The backup options that can be set for the standard version are also available with the cluster-aware version.

For information about backing up data in the Exchange Server Clustered environment, Oracle RAC setup, SQL Server Failover Cluster, and MySQL Server Failover Cluster, see the relevant plug-in user’s guide.

Note the following:

•

When you back up a virtual client, the data is backed up from a single client and it is accessed from the controlling node.

•

When backing up cluster data using a virtual client, only the LUNs owned by the active node are backed up; the LUNs owned by passive nodes are not backed up.

•

In an Active/Active cluster setup, the NetVault Server may start a backup on a secondary node even if the “primary only” option is selected for the backup. In such cases, the backup is redirected to the primary node, but the server only communicates with the secondary node to which it sent the message to start the backup. If the secondary node becomes unavailable while the backup is in progress, the server loses communication with the parent process running on this client. This process is not restarted even if the secondary node becomes available immediately. Consequently, the backup job becomes unresponsive.


	NOTE: For cluster backups and restores, the virtual client name is displayed on the Job Status page and the actual client name is displayed on the View Logs page.

Cluster failover during backups

If a failover occurs during a backup, the job is aborted and the status “Job Failed” is returned. You can use the Job Retry scheduling option to run the job again after the failover completes.

With the Plug-in for FileSystem, when a failover occurs on Windows, the reason for the failover has a direct bearing upon the status of the failed job. Depending on the reason, the job returns the following exit status:

•

If the machine in control of the cluster resources goes offline during a backup, the job is aborted and the status “Job Failed” is returned. You can use the job retry feature to run the backup again.

•

If the machine in control of the cluster remains online but the actual cluster resource that is being backed up fails, the job is aborted and the status “Backup Completed with Warnings” is returned. The scheduled job retries do not work for such jobs. In this scenario, examine the job logs to find the missing data or run the job again to back up the data.

Restores using cluster-aware plug-ins

The procedure for performing restores using the Plug-in for FileSystem is similar for both standard and virtual clients. The backups are restored from the virtual client node and not the actual client node. When you submit a restore job, the plug-in communicates with the cluster service to determine the controlling node and targets this machine for the restore.

For information about restoring data in the Exchange Server Clustered environment, Oracle RAC setup, SQL Server Failover Cluster, and MySQL Server Failover Cluster, see the relevant plug-in user’s guide.


	NOTE: For cluster backups and restores, the virtual client name is displayed on the Job Status page and the actual client name is displayed on the View Logs page.