Why a Hyper-V Checkpoint Isn’t a Backup

Hyper-V checkpoints, formerly called Hyper-V snapshots, are a convenient way to place a virtual machine back to an earlier state. You can use them for development and testing, for example, when you need to have an exactly identical starting point for your tests. Because the checkpoint functionality acts like it’s restoring a VM, many beginners misinterpret this functionality to function like a backup. But it does not.

Because hard drive failures are common and because checkpoints are stored on the same hard drive or disk array as the VMs, the protection potential of a checkpoint is very limited. Yes, it can protect against software corruption, for example, one that is caused by a malfunctioning Windows Update. However, there are many scenarios that lead to a broken VM and checkpoints do not cover many of those scenarios.

While Windows Server 2012 R2 and newer versions of Hyper-V allow checkpoints to reside on a separate hard drive, those files are still interdependent and interrelated. So even if your checkpoints are on different drives, the main issue with checkpoints is still the same. Hence, it’s impossible to get the VM up and running if any of those hard drives should fail.

Apart from the fact that Hyper-V checkpoints are negatively affecting the performance of the Hyper-V host, Microsoft recommends against using them on production systems for several reasons. First is the already mentioned significant performance penalty involved with snapshots. Second, checkpoints aren’t portable like standalone VHDs. They depend on CPU and hardware architecture. And third, Microsoft sometimes introduces bugs related to snapshot handling and you wouldn’t want to add that risk on top of all hard drive failure risks. For example, when a checkpoint is merged live, there is a potential for the VM to become corrupted, if the system hangs or crashes, or the disks run full, for example.

A RAID array isn’t going to help much with data corruption risks resulting from OS software bugs either. In addition, by the time it is discovered that a hard disk issue is present, the damage is already done and replicated by the RAID. Snapshots are more likely to cause hard drive failure, too, due to the higher disk I/O they cause. The wear on the hard drive is hence greater.

A Hyper-V Backup on the other hand takes a copy of the entire VM, with checkpoints and everything else related to the VM. It’s important to use a separate media for your backups, such as an external hard drive connected via USB, a network share, a NAS device, or a remote FTP server, for example. For example, it would make sense to copy your VMs to another isolated machine, which isn’t connected in any way with the original.

Think about electric discharge, electric surges, virus damage, software corruption, and mechanical failures due to overheating or wear and tear, are likely to affect the entire server. A RAID array is also likely to break with several disks damaged simultaneously since they are exposed to the same conditions over an entire lifetime. If your Hyper-V backup is stored on a separate device, chances are it won’t be affected by all of the above. Hyper-V backup is different from Hyper-V checkpoints in that it isolates the VM and clones it in its entirety on a separate system.

While you can work productively without snapshots and checkpoints, you can’t without solid, reliable, and frequent Hyper-V backups, since backups are the only insurance you have against data loss.

Check out our backup solution for Hyper-V and back up your Windows Server as well as your virtual machine guests with a single, lightweight, and affordable solution.