→ Applies to: SynetoOS 4.x, SynetoOS 5.x

When protecting virtual machines, you can choose between two types of recovery points: Crash-Consistent and Application-Consistent. Both serve different purposes, and understanding their differences can help you decide which is best for your needs.

Crash-Consistent Snapshots

Crash-Consistent snapshots are created instantly at the storage level, which means that no data is copied, only the blocks used in the metadata are recorded. This type of snapshot is completely independent of the operating system or applications running on the virtual machine.

A major advantage of Crash-Consistent snapshots is that they can be taken very frequently (even every 1, 5, 10, or 15 minutes), without causing any performance impact on the virtual machine or host. This makes them ideal for situations where frequent restore points are critical.

However, while these snapshots ensure that data is captured at a specific point in time, they do not guarantee that the application data is in a consistent state. If a database or other transaction-intensive application is running at the time of the snapshot, there is a risk of data corruption or incomplete transactions when it is restored.

Application-Consistent Snapshots

Application-Consistent snapshots ensure that the data inside the virtual machine is in a valid, recoverable state. These snapshots are created using the VMware API, using VMware Tools to temporarily pause (quiesce) the virtual machine's file system.

For Windows-based virtual machines, the Volume Shadow Copy Service (VSS) is used to ensure that applications such as databases, Active Directory, or email servers are in a consistent state before the snapshot is taken. This means that when you restore from an application-consistent snapshot, the applications will be able to resume without data corruption or incomplete transactions.

One important thing to keep in mind is that Application-Consistent snapshots have a performance impact. Each snapshot causes a temporary pause (also called an application stun) that can slow down the virtual machine and host, especially if snapshots are created too frequently. Although snapshots can be created as frequently as every 15 minutes, it is important to monitor system performance, user feedback, and resource usage to find the right balance.