Changed Blocked Tracking for Microsoft Hyper-V
Enable Changed Block Tracking (CBT) to optimize incremental backups of virtual machines by reading only the allocated and modified portions of virtual disks. Incremental backups can use SnapProtect CBT, native Microsoft CBT, or Cyclic Redundancy Check (CRC):
- For Hyper-V virtual machines running on Windows Server 2012 R2 editions, SnapProtect CBT enhances backup performance of VHDX-format disks.
The SnapProtect CBT mechanism is provided as a CVCBT driver along with the Virtual Server Agent (VSA) software. If CBT is enabled and if both the VSA proxy and the virtual machine proxy are running on a Windows Server 2012 R2 node, SnapProtect changed block tracking is used. The driver is supported on all editions of Windows Server 2012 R2 (Hyper-V Core, Standard and Datacenter editions). The VHDX files can reside on local volumes, cluster shared volumes, or SMB shares. CBT is not supported for backups on remote SMB (Server Message Block Protocol) shares.
- A Hyper-V cluster using Windows Server 2016 uses native Microsoft changed block tracking. For each virtual machine, the backup takes a checkpoint that includes only that virtual machine. Once the backup is completed, the checkpoint is converted to a reference point with a unique ID and saved with the backup. Changed block tracking uses the reference point to identify changes since the last backup. The reference point is retained until the next backup is completed and a new checkpoint is generated. In a mixed mode cluster where some nodes have not yet been upgraded from Windows Server 2012 R2 to Windows Server 2016, Windows Server 2016 operations are not supported until all nodes in the cluster are upgraded to 2016 and the functional level of the cluster is switched to Windows Server 2016 mode. Microsoft CBT is not available in a mixed mode cluster or for a Windows Server 2016 nodes that includes older version guest VMs (prior to version 8). For more information, see Windows Server 2016 Support.
- Backups of VHD-format disks and virtual machines running on Windows Server 2008 R2 or Windows Server 2012 continue to use the Cyclic Redundancy Check (CRC) mechanism to back up the parts of virtual disks that have changed since the last full backup. With CRC, the amount of data that is transferred to the MediaAgent is less, but the backup operation must read and compute the CRC checksum against all disk blocks. The time required for the disk I/O operations increases the total time of incremental backups to nearly as long as full backups.
For instructions on enabling or disabling CBT, see Enabling or Disabling Changed Block Tracking.
How SnapProtect CBT Works
The CVCBT driver is available with the software on each Hyper-V node in a cluster. After enabling CBT backups, when you run the first backup of a CBT enabled subclient, the CVCBT driver is activated and starts tracking the changed blocks on the virtual disks for the given virtual machines. The CBT driver monitors the virtual disks and maintains a bitmap (.cvf) file of the blocks that have changed. During subsequent incremental backups, these bitmap files are read and only the changed data blocks are backed up.