Backup

Recommendations for the backup of vRealize Operations are listed here.

• It is highly recommended to take only backups during quiet periods

  Since a snapshot-based backup happens at the block level, it is important that they are limited, or no changes being performed on the cluster configuration. This helps to ensure a healthy backup.

• It is best to take the cluster offline before backups

  This ensures the data consistency across the cluster and internally within the nodes. If the VM cannot be powered off, you can either shut down the VM before the backup or disable quiescing.

• Do not quiesce the file system when the cluster remains online

  If the cluster remains online, backup your vRealize Operations multi-node cluster by using vSphere Data Protection or other backup tools, disable quiescing of the file system. Snapshots with quiesce activated is unsupported and may cause problems when restoring.

• Use resolvable host names and static IP addresses for all nodes

  The hostname must be resolvable to ensure a consistent communication between nodes. If the hostname fails to resolve or the IP has changed, problems may result.

• All nodes must be powered off and accessible during backups

  All nodes in the cluster must be in the same powered state when taking backups to maintain a consistent state when restored. If nodes cannot be powered off, disable quiescing.

• Backup the entire cluster to include all VMs

  Restoring only part of the cluster is unsupported and may cause synchronization problems preventing the cluster from going online.

• All VMDK files that are part of the virtual appliance must be backed up

  Include all VMDK files in the backup; otherwise, the node may not properly connect to the cluster when restored.

• Backup of all nodes must be performed at the same time

  Initiate backups of all nodes (primary, replica, data, witness, and cloud proxy) at the same time to maintain the synchronization across nodes. Each node may complete their backup at a different time but starting the backup process at the same time minimizes the time differential between nodes when restored.

• Perform backups outside of vRealize Operations internal operations. By default, the following processes run:

  • Dynamic Threshold (DT) Calculation at 2:00 am

  • Capacity Calculation (CIQ) at 9:00 pm (vRealize Operations 6.6.1 and earlier)

  • predictive Distributed Resource Scheduler (pDRS) at 6:00 pm

  • Cost Calculation at 9:00 am (introduced in vRealize Operations 6.7)

  Avoid processing overhead of the cluster by performing backup when DT, CIQ, pDRS or Costing are not running. These default times can be modified to avoid conflicts during backups

• Backup at different times from infrastructure backups

  If there is a process which maintains a separate backup of the infrastructure, avoid taking backups of the vRealize Operations cluster at the same time.

• Do not backup cloud proxies if they are already removed from the vRealize Operations cluster

  Remove backing up cloud proxies if they have been removed from the vRealize Operations cluster to prevent the cluster confusion when restored.