Are you getting confused (like everyone else) about checkpoints, snapshots, VSS snapshots, Hyper-V snapshots, and SAN snapshots?
It is confusing thanks to Microsoft and a couple of others….
Let’s get to work:
Hyper-V Snapshots are Now Called Checkpoints
Hyper-V snapshots were called snapshots in Windows Server 2008 and Windows Server 2008 R2.
Then Microsoft renamed them “to prevent confusion” to Hyper-V Checkpoints.
Hyper-V snapshots are essentially the same thing as VMware snapshots and VirtualBox snapshots.
The state of a virtual machine is saved, creating a restore point so you can go back in time and quickly restore the hard drive and actual VM state of a particular point in the past, when the snapshot was created.
Hyper-V checkpoints are for demo and test purposes only. They are not meant to be used on a production system.
This section describes potential issues with checkpoints.
This ‘going back in time’ feature sound good and useful but you need to know what you are doing. First, some transaction and replication-based services and applications may fail big time if you ‘set the clock back’. If the internal state goes back to a prior state, but the data has been already replicated elsewhere…there will be problems on the horizon.
The quality of snapshots and their management depend solely on the virtualization platform used. It’s not unusual for snapshots to end up becoming unbootable or even completely disappear from the system. There are bugs in Windows, in Hyper-V, in VMware, etc. that haven’t been fixed yet and that can lead to data loss scenarios. It’s always a good idea not to blindly trust and deploy those features without have a good backup first.
Loss of Links:
Obviously–or not immediately as obvious–the VM will lose current internet and network connections if reverted back to the past. It’s like beaming yourself to a different place when you don’t expect it. Client TCP links for example will need to be reset when a VM is reverted to an old state and the service application needs to be capable of handling that, too.
Virtual differencing disks are being used to accomplish snapshot functionality. Most of the time these disks are also dynamically expanding. A big no-no on production systems. You can easily clog your server with a couple of these snapshots. Expect a performance degradation of 300 to 1000% when using snapshots versus a single, fixed sized VHDX.
Volume Shadow Copy Service (VSS) Snapshots, aka. Shadows
VSS snapshots (shadows) are point-in-time background images of hard drives used by the Volume Shadow Copy Service in Windows.
These snapshots are necessary when you want Hyper-V live backups to take place. VSS is a framework that allows services to register and prepare for live backup. Without VSS there wouldn’t be a way to run backups while Windows (and services) are running and ensure the backup is application and crash consistent.
VSS snapshots have absolutely nothing to do with Hyper-V snapshots and checkpoints.
A bad performing server may have dozens of shadows in the system. As with Hyper-V virtual machine checkpoints, VSS shadows can case an enormous resource spike and performance degradation, even when no backups are running. Because every hard disk block access needs to be virtualized and potentially rerouted, this backup technique adds considerable stress to your servers. You would definitely want to check you don’t have orphaned VSS shadows on your servers. In addition, you may want to consider turning off System Restore and not using Windows Server Backup, as these technologies create lots of persistent shadows periodically.
SAN snapshots are a technique similar to VSS snapshots, and may in fact be coupled to VSS, such that VSS shadows are delegated to a SAN device. A hardware-based implementation of a disk shadow is performing far better; however, you will find that there are plenty of compatibility issues from SAN to SAN and VSS. You’ll want to check your file backup software to make sure it works with the SAN drivers you have.
Snapshots aka checkpoints, VSS shadows and SAN snapshots are all great tools; yet, they should be used with caution. Ideally the IT admin should be aware of potential issues that may emerge from using them.
As always, it’s always a good idea to use a good backup solution for Hyper-V and VMware and protect your hard work!