For best results, it is recommended that you plan your strategy for disaster recovery before an asset fails. Following is a high-level overview of the steps you must complete to implement unified bare metal protection for your Windows assets. It identifies steps to complete before and after an asset fails.
Perform the following before an asset fails
Step 1: | Determine which Windows operating system the asset is running and whether it is the 32-bit or 64-bit version. For instructions, see the Microsoft document Which Windows operating system am I running? |
Step 2: | Verify that the asset’s operating system is supported and review the Prerequisites for unified bare metal recovery of Windows image-level backups. |
Step 3: | Run image-level backups that include all critical system volumes, as described in To create an image-level backup job. |
Step 4: | (Optional, recommended) Perform a test recovery. For details, see Considerations for performing a test unified bare metal recovery. |
To recover a failed asset
Step 5: | Perform unified bare metal recovery (UBMR) using the procedures in Performing unified bare metal recovery. |
Consider the prerequisites for UBMR as you plan your disaster recovery strategy.
Requirement |
Description |
||||||||||||||||||||||||
---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
Operating systems |
Recovery to identical physical hardware and virtual machines is supported for the operating systems listed below. Recovery to dissimilar hardware is NOT supported. Supported client operating systems:
Supported server operating systems:
|
||||||||||||||||||||||||
The backup used for recovery must meet these requirements:
Notes:
|
|||||||||||||||||||||||||
UBMR ISO image |
For the recovery, you must use the 64-bit UBMR ISO image provided on the your Unitrends appliance. The ISO contains WinPE (a minimal version of Windows used for installations) and the Unitrends Unified Bare Metal Recovery wizard that guides you through the recovery. To prepare for DR, it is recommended that you do the following:
For details, see " Unified bare metal recovery of image-level backups". |
||||||||||||||||||||||||
Firmware interface type |
Supported for BIOS- and UEFI-based assets. The firmware interface type (BIOS or UEFI) of the recovery target machine must match that of the failed asset. |
||||||||||||||||||||||||
Disk configuration |
|
||||||||||||||||||||||||
Network adapter |
Wireless network adapters cannot be used for the recovery. |
||||||||||||||||||||||||
Drivers |
Drivers needed for UBMR are determined by the operating system and hardware of the asset you are recovering and whether you are recovering to a physical or virtual machine. You might need to add drivers during different stages of the recovery.
|
||||||||||||||||||||||||
Processor features on the recovery target |
WinPE 5.0 requires that these processor features are enabled on the recovery target machine: NX, PAE, and SSE2. Ensure that these features are enabled on the recovery target machine before booting from the ISO image. For instructions, see KB 1190. Machines that do not have these processor features cannot be used for the recovery. |
||||||||||||||||||||||||
Use the procedures in Performing unified bare metal recovery for the following recovery scenarios:
The virtual host must support the operating system (OS) of the Windows asset you are recovering. (See the VMware, Microsoft, or Citrix documentation for details. For Hyper-V, see this Microsoft article: Should I create a generation 1 or 2 virtual machine in Hyper-V?) For example, you cannot recover Windows 2016 to ESXi 5.1 or Hyper-V 2008 R2. |
Use the following procedures to perform unified bare metal recovery. For a successful recovery, be sure to run the procedures in order. Before you start, it is recommended that you read Prerequisites for unified bare metal recovery of Windows image-level backups for prerequisites and supported recovery scenarios.
For the recovery, you must use the 64-bit UBMR ISO image provided on the Unitrends backup appliance.
To access the unified bare metal recovery ISO image
1 | Mount a working asset to the virtual_failover share on your Unitrends appliance by entering \\<ApplianceIP>\virtual_failover in the Windows File Explorer. |
2 | Downloadwinbm.iso, the ISO for 64-bit assets. |
3 | Do one of the following: |
• | If recovering to a physical machine, burn the ISO to CD. Refer to the documentation for the burner you are using to walk you through the process of creating a bootable CD from an ISO image, which is not the same as burning an ISO image to a CD. |
OR
• | If recovering to a virtual machine, save the ISO in a location that you can access from your hypervisor. |
4 | Proceed to Step 2: Prepare the recovery target machine. |
You can recover a failed asset to a physical or virtual machine (VM). Use one of the following procedures to prepare the recovery target machine:
Note: WinPE 5.0 requires these processor features to be enabled on the recovery target machine: NX, PAE, and SSE2. Be sure these features are enabled before you boot from the ISO image. For instructions, see KB 1190. Machines that do not have these processor features cannot be used for the recovery.
1 | Verify that the machine’s hardware and firmware interface type (BIOS or UEFI) match that of the failed asset. Recovery to dissimilar hardware is not supported. |
2 | Make sure the machine has enough disk space for the recovery. |
• | You must recover to destination disks that are the same size as the original disks or larger. Recovery fails if the disk does not have enough space. |
• | Any existing data on the destination disks is overwritten or deleted during the recovery, even if the disk has more than enough space. Volumes on the target disk that were excluded from backup may also be overwritten. Before performing a recovery, make sure you have additional copies of any data on the destination disk. |
3 | Load the disk containing the bootable ISO image into the machine’s CD/DVD drive. (For details on creating the ISO disk, see Step 1: Access the unified bare metal recovery ISO image.) |
4 | Proceed to Step 3: Run the Unified Bare Metal Recovery wizard. |
A VMware, Hyper-V, or XenServer virtual machine can be used as the recovery target. You can create a new VM or edit the settings of an existing VM. Prepare the recovery target VM to meet these requirements:
• | Make sure the VM’s firmware interface type (BIOS or UEFI) matches the firmware interface type of the failed asset. |
• | Add enough memory to satisfy Microsoft’s support guidelines for the operating system being recovered. The recovery requires at least 1 GB of memory. |
• | Assign the VM a virtual hard disk with enough space for the recovery. You must recover to destination disks that are the same size as the original disks or larger. Recovery fails if the disk does not have enough space. |
• | Any existing data on the destination disks is overwritten or deleted during the recovery, even if the disk has more than enough space. Volumes on the target disk that were excluded from backup may also be overwritten. Before performing a recovery, make sure you have additional copies of any data on the destination disk. |
• | Windows 10 and Windows Server 2016 VMs – The Windows Secure Boot option is not supported. Ensure that Secure Boot is set to OFF before you perform the bare metal recovery. |
• | For a VMware virtual machine, make sure you are using the E1000 NIC. This NIC is only required during the recovery. After rebooting the recovered VM, you can use a different NIC. |
• | Edit the VM settings to boot from the bare metal ISO. (For details on accessing the ISO, see Step 1: Access the unified bare metal recovery ISO image.) Procedures vary by hypervisor. A VMware example is given here: |
1 | Power off the VM. |
2 | Right-click the VM and select Edit Settings: |
3 | On the Hardware tab, click Add, select CD/DVD Drive and click Next: |
4 | Select Use ISO image and click Next: |
5 | Browse or enter a path to select the unified bare metal recovery ISO, check the Connect at power on box, and click Next: |
6 | Click Next to continue: |
7 | Click Finish to add the device: |
8 | Click OK to save: |
9 | Do one of the following: |
• | If the target VM does NOT have an operating system installed, proceed to Step 3: Run the Unified Bare Metal Recovery wizard. |
OR
• | If the target VM has an operating system installed, you must boot into BIOS and change the boot order to force the VM to boot from the bare metal ISO. See the remaining steps in this procedure for a VMware example. |
Note: The remaining steps provide an example of how to boot into BIOS and change the VM's boot order. This example is run on VMware. Details will vary by hypervisor.
10 | Ensure that the VM is powered off. |
11 | Right-click the VM and select Edit Settings. |
12 | On the Options tab, select Boot Options, check the Force BIOS Setup box, then click OK: |
13 | Power on the VM. The VM boots into BIOS. |
14 | Launch the VM console. |
15 | On the Boot menu in the BIOS Setup Utility, select CD-ROM Drive and press + to move it to the top of the list. Press F10 to save. |
16 | Press Enter to save and exit. |
17 | Power off the VM. |
18 | Proceed to Step 3: Run the Unified Bare Metal Recovery wizard. |
Perform the recovery by using the Unified Bare Metal Recovery (UBMR) wizard. See the following topics for details:
Before an asset fails, you can perform a test unified bare metal recovery without impacting the original asset. As long as you assign the recovered asset a unique IP address and rename it, the test recovery does not cause any network conflicts with the original asset. IP conflicts will occur if you perform a test recovery and do not assign the recovered asset a unique name and IP address.
To perform the test recovery, use the Perform the recovery by running the Unified Bare Metal Recovery wizard procedures.
Use the procedures in this section to perform the recovery by running the UBMR wizard. These procedures apply to any supported target machine (a physical machine or a virtual machine).
Notes:
• | You must prepare the recovery target machine before running the UBMR wizard. (See Step 2: Prepare the recovery target machine for details.) |
• | If you exit the UBMR wizard before you are finished with the recovery, you are taken to a command window. To return to the wizard from this window, run the following command: z:\pcpb\Restore.exe. |
For a successful recovery, you must run the procedures in the following order:
These steps ensure that the recovery target machine can communicate with the appliance that is storing the backup you will use for the recovery.
1 | Boot the recovery target machine from the bare metal ISO image. The machine boots into WinPE and launches the Unified Bare Metal Recovery wizard. |
Notes:
• | If you see a message stating that you must set up networking to continue or that no disks are detected on the local system, you might need to load drivers into WinPE. Click OK to allow the boot to continue. You can load drivers later in this procedure. |
• | If you are recovering to a VM and the Windows login screen displays, you must boot into BIOS and change the boot order to force the VM to boot from the bare metal ISO. For a VMware example, see step 10 - step 17 of the To prepare a VM for unified bare metal recovery procedure. |
2 | Select a Network Adapter. If the machine has more than one adapter, the default adapter displays first. |
If a network adapter does not display, WinPE cannot detect one. To resolve this issue, perform the following:
• | Ensure that the network cable is plugged in to an active port. |
• | If the adapter is connected to the network and WinPE is unable to detect it, you must load a Windows 8 network driver into WinPE. In the Load WinPE Drivers section, specify a driver by browsing or entering a path, then click Load Driver. |
3 | Review network settings and modify as needed. |
• | If DHCP is configured for your network, network settings are assigned automatically. |
• | If DHCP is not configured, or if you want to configure network settings for the target machine manually, click Change Settings. Then enter a unique IP address, the Subnet Mask, and the Gateway. Network settings do not need to match those of the original asset. The only requirement is that the machine can communicate with the appliance that is storing the backup you will use for recovery. |
Note: The network settings that you configure during this step are used only for the recovery. They are not applied to the network adapter when you reboot the recovered operating system. Before connecting the recovered asset to your network, you will reconfigure the asset’s network settings.
4 | Select the time zone of the Unitrends appliance storing the backup or hot backup copy that you will use for recovery. |
5 | Click Next, then proceed to Step 2: Select an appliance and the backup to recover. |
1 | For Backup Source, specify the Unitrends appliance that is storing the backup or hot backup copy you will use for the recovery. You can either: |
• | Select the appliance from the Backup Source list. This list contains all appliances on the same subnet as the recovery target machine. |
• | Enter the IP address of an appliance on a different subnet. |
2 | Enter the Username and Password credentials of a UI user account that has superuser or administrator privileges. Click Login. |
3 | Select the Windows asset from the Client list. |
Note: Assets that have eligible image-level and/or file-level backups display in the list. (For more on eligible backups, see An eligible image-level backup.)
4 | Select a recovery point in the calendar. |
• | Days with a backup display in bold on the calendar. |
• | If multiple backups exist on a given day, the different times for these backups display in the Recovery Point list. |
• | If the recovery times do not match the backup times on the Unitrends appliance, verify that you selected the time zone of the appliance that you are using for the recovery (in the Step 1: Boot into WinPE and set up the local environment procedure above). |
• | The selected Recovery Point shows the backup date/time, type ( Img for image-level or File for file-level) and mode (Full or Incr for incremental). |
5 | Click Next, then proceed to Step 3: Map drives. |
After selecting an appliance, asset, and recovery point, you must map the failed asset’s disks to disks on the recovery target machine.
1 | On the Disk Mapping screen, select the disks to recover. |
• | If the recovery target is similar to the original machine, the wizard automatically maps the backed up disks to the destination disks. |
• | You must select system critical disks. Other disks are optional. |
Note: It is recommended to select all disks. The only way to recover all data on a disk is by selecting the disk during this mapping step. After completing the bare metal recovery, you can opt to recover individual files from the backup (as described in Recovering files from Windows image-level backups), but you cannot recover entire disks or volumes.
• | The wizard does not require that you select all critical disks to proceed, but for the recovery to succeed, you must select the disks needed to boot the operating system. If you are unsure which are required, select all disks. |
• | To perform the mapping manually, uncheck the Restore to original system (automatic mapping) box, then manually add disks to the Destination Disks area at the bottom of the screen by highlighting the destination disk and clicking Add. (To remove a disk, select the disk and click Remove.) |
Consider the following when selecting a destination disk:
• | Any existing data on the destination disks is overwritten or deleted during the recovery, even if the disk has more than enough space. Volumes on the target disk that were excluded from backup may also be overwritten. Before performing a recovery, make sure you have additional copies of any data on the destination disk. |
• | You must recover to destination disks that are the same size as the original disks or larger. Recovery fails if the disk does not have enough space. |
• | Disks are assigned numbers during recovery that may not match the numbers from the original disks. |
• | Volumes are assigned letters during recovery that may not match the letters from the original volumes. |
2 | Click Next. |
3 | Click Yes to confirm. |
4 | Proceed to Step 4: Start the recovery. |
Run one of the following procedures to start the recovery:
1 | Click Restore. Proceed to the next step when the job completes. |
Note: The recovery must complete within 24 hours. If many jobs are running on the appliance, move the UBMR job ahead in the queue by canceling running jobs or pausing queued jobs. For details, see Managing active jobs in the Administrator Guide for Recovery Series, Recovery MAX, and Unitrends Backup.
The recovery can take some time. Monitor job progress in the Restore Progress Job area at the top of the screen. Sample messages are given here:
2 | Click OK to close the Status message. |
3 | Click Reboot to reboot the target machine. |
After the target machine reboots, proceed to Step 4: Complete the unified bare metal recovery. (If the target fails to boot, you need to inject drivers. Proceed to step 4 for details.)
4 | If the target fails to boot, inject storage drivers: |
• | Boot the machine from the ISO disk. |
• | After booting from the ISO disk, you are returned to the first screen of the wizard. Click the Driver Injection button on this screen. |
Note: Do not use the Load Driver button under Load WinPE Drivers. This loads drivers into WinPE. Because your operating system has already been recovered, you must inject drivers into the recovered operating system instead.
• | Browse to select the driver and click Done to add it to the list of drivers to inject. Repeat as needed to add more drivers to the list. |
• | Highlight the volume containing the operating system files in the asset’s Offline Disk and Volume(s) Information area. Then highlight a driver and click Inject. Repeat this step as needed to inject all the necessary drivers. |
5 | When you receive a message stating that the driver injection was successful, you are ready to reboot the recovered asset. Remove the bootable ISO CD, then click Reboot. |
6 | After the recovered asset boots, proceed to Step 4: Complete the unified bare metal recovery. |
Note: If the asset fails to boot, you might need to add additional drivers.
1 | Click Restore. Proceed to the next step when the job completes. |
Note: The recovery must complete within 24 hours. If many jobs are running on the appliance, move the UBMR job ahead in the queue by canceling running jobs or pausing queued jobs. For details, see Managing active jobs in the Administrator Guide for Recovery Series, Recovery MAX, and Unitrends Backup.
The recovery can take some time. Monitor job progress in the Restore Progress Job area at the top of the screen. Sample messages are given here:
2 | Click OK to close the Status message. |
3 | Inject an ESX, Hyper-V, or XenServer guest storage driver by doing these steps: |
Notes:
If you attempt to reboot without adding the driver, the VM boots to a blue screen. Return to the wizard by booting from the bare metal ISO image. The first wizard screen displays. Click Driver Injection. Then inject the necessary driver as described below.
• | Click Add Driver. |
• | Select the guest storage driver. Click Done to add the driver to the list of drivers being injected. |
• | Highlight the volume containing the operating system files in the asset’s Offline Disk and Volume(s) Information area. Then highlight the guest driver and click Inject. |
• | Click Yes to continue. |
• | Click OK to close the Restore message. |
4 | When you receive a message stating that the driver injection was successful, you are ready to reboot the VM. Edit the VM settings so it no longer boots from the bare metal ISO. Next, click Reboot then click Yes to continue. |
5 | After the VM boots, proceed to Step 4: Complete the unified bare metal recovery. |
Use these steps to complete the recovery:
1 | Configure network settings for the recovered asset. The network settings that were used for the recovery are not retained after booting into the recovered operating system. Consider the following when configuring network settings: |
• | If the original asset is still connected to the network, you must assign the recovered asset a unique IP address and rename it before connecting to the network to avoid conflicts. |
• | If the original asset is no longer connected to the network, you can assign the recovered asset the same IP address as the failed asset. |
• | If you are using DHCP and you added the original asset to the backup appliance by using only the asset’s name, the appliance discovers the recovered asset after you connect it to the network. If the original asset is still connected to the network, rename the recovered asset to prevent conflicts. |
• | If you configure the recovered asset to use the same name and IP as the original asset, the appliance treats the recovered asset as if it is the original asset. No changes are needed on the Unitrends appliance. |
• | If you configure the recovered asset to use a different name and IP than the original asset, the appliance treats the recovered asset as new asset. To protect the recovered asset, add it to the appliance and add or modify job schedules. |
2 | Do these additional steps as needed: |
• | For Hyper-V servers — After recovery, you must run the following command on the Hyper-V server: bcdedit /set hypervisorlaunchtype Auto. Then reboot the server. |
• | For Hyper-V servers recovered to a VMware or XenServer virtual machine — If the Hyper-V server had Integration Services installed, you must remove the Integration Services provider from the recovered VM before running Unitrends backups, as described here: |
– | Log in to the recovered VM as administrator and open a command prompt. |
– | Enter this command: vssadmin list providers |
– | In the output, find Hyper-V IC Software Shadow Copy Provider and note the Provider ID. Example ID: {74600e39-7dc5-4567-a03b-f091d6c7b092} |
– | Using a registry edit tool like regedit, remove the provider key and all subkeys: [HKEY_LOCAL_MACHINE\SYSTEM\CurrentControlSet\Services\VSS\Providers\{74600e39-7dc5-4567-a03b-f091d6c7b092}] |