Consider the following as you plan for disaster recovery of your Hyper-V virtual machines.
The following Unitrends software requirements must be met to perform Hyper-V instant recovery:
• | The Unitrends appliance must be running release 8.0 or higher. |
• | Unitrends Windows agent release 8.0 or higher must be installed on the Hyper-V server hosting the protected VM and on the Hyper-V server used as the restore target. |
• | Unitrends recommends reloading the list of VMs on the appliance after installing the agent on the Hyper-V server hosting your VMs. To reload the list of VMs, select the Hyper-V application in the Navigation pane, click Backup, then click the reload icon in the lower right corner of the VMs to Protect list. |
• | Backups taken with a pre-8.0 agent can be used as long as agent release 8.0 or higher has been installed on the server hosting the VMs. This agent is required for the appliance to retrieve generation information for the VMs. |
The following Hyper-V server requirements must be met to perform Hyper-V instant recovery:
• | The restore target must be one of the following: |
– | A Windows Server with the Hyper-V role enabled, running Windows 2012 or a higher version listed in the Compatibility and Interoperability Matrix. |
– | A Hyper-V Server running 2012 or a higher version listed in the Compatibility and Interoperability Matrix. |
• | The restore target can be the original Hyper-V server or an alternate Hyper-V server. |
• | The restore target must be running the same version as the Hyper-V server hosting the original VM, or a higher version listed in the Compatibility and Interoperability Matrix. It is recommended that you restore to a Hyper-V server whose version matches that of the original, where possible. |
Note: Backups from older versions of Hyper-V can be used for Hyper-V instant recovery as long as the target server is 2012 or higher, and the Unitrends appliance and Windows agent on the Hyper-V host are running version 8.0 or higher.
• | The target server must be registered to the Unitrends appliance from which you are performing the instant recovery. |
• | The target Hyper-V server must support the operating system (OS) of the VM you are recovering. (See the Microsoft documentation for details.) For example, you cannot recover a Windows 2016 VM to Hyper-V 2008 R2. |
• | You can perform instant recovery using local or replicated backups. |
The following Hyper-V VM requirements must be met to perform Hyper-V instant recovery:
• | All disks created on the Hyper-V virtual machine must have unique names. |
• | Since Unitrends does not create clones, the restored VM is configured with the latest hardware version that is supported by the target Hyper-V server. |
• | The restored VM's configuration version is the highest version supported by the target Hyper-V server, up to version 5.0. If the target server supports versions higher than 5.0, the VM is created with version 5.0. |
Note: Starting with VM configuration 6.2, Microsoft implemented a binary-based configuration format (VMCX). Due to this change, different recovery methods are required, which will be available in a future Unitrends release.
• | Disks excluded during backup, including independent and physical RDM disks, are not restored with instant recovery for Hyper-V. |
• | Hyper-V Instant Recovery is not supported for Host component (AzMan Security Database for Hyper-V) backups of the original virtual machine. These are not bootable backups. |
Hyper-V instant recovery supports clustered and non-clustered virtual machines. The following requirements must be met to recover a clustered VM:
• | The target Hyper-V server determines the cluster status of the new virtual machine. To create a clustered VM for instant recovery, you must select a cluster as the target Hyper-V server. If you select an individual member node, the resulting VM is not clustered. |
• | When creating a clustered VM for instant recovery, you must select the network switch common to all nodes in the cluster. If you do not select this switch, the new VM will lose network connectivity if it fails over to another cluster node. |