What is Snapshots?
Learn how instance restore points and volume snapshots copy your disks, how you restore them, what they keep in use, and how they're billed.
Snapshots are point-in-time copies of your disks. An instance restore point copies every disk of an instance in one step. A volume snapshot copies one volume. You restore either kind by creating a new instance or a new volume from it.
Key features
- Instance restore points. Copy the OS disk and the data disks of an instance at once, while the instance keeps running. See Create an instance restore point.
- Volume snapshots. Copy one volume while it stays online and writable, for example before an upgrade. See Create volume snapshots.
- Restore to new resources. Create a new instance from a restore point, and a new volume or a new instance from a volume snapshot. See Restore from a snapshot.
- Lineage. Link the snapshots of a volume into a chain to record their order. See Snapshot chains and lineage.
How it works
The two kinds of copies differ in what they copy and in how you use them:
| Instance restore point | Volume snapshot | |
|---|---|---|
| What it copies | Every disk attached to the instance: the OS disk and the data disks. | One volume. |
| Where it's listed | ComputeInstance snapshots | StorageVolume snapshots |
| How you restore it | Create a new instance from it. The instance gets a new copy of every disk. | Create a new volume from it, or create an instance that boots from it. |
| What it keeps in use | The instance's volumes can't be deleted while the restore point exists. | Its source volume can't be deleted while the snapshot exists. |
| Chains | None. | Optional, with Link to previous snapshot. |
Restoring never changes the original. There's no way to roll an existing instance or volume back in place: you always get a new instance or a new volume, and you then move your workload or your data to it.
In the creation wizard, both kinds appear under the Snapshot boot source, marked VM snapshot or Volume snapshot. The console also calls restore points VM snapshots.
A restore point holds its disk copies in the cloud as volume snapshots, but they don't appear in the volume snapshot list. They're deleted when you delete the restore point.
Use cases
- Before a change. Take a restore point or a volume snapshot before an upgrade, a migration, or a configuration change, so you can recover the earlier state on a new instance or volume.
- A copy of an environment. Create a new instance from a restore point to test a change on a copy of a production instance.
- Recover files. Create a volume from a volume snapshot, attach it to an instance, and copy the files that you need.
Billing
Snapshots bill these dimensions:
| Dimension | How it's billed |
|---|---|
| Instance restore points | As image storage, per GB-month of the restore point's allocated size, while the restore point exists, when the region has a price for it. |
| Volume snapshots | As volume snapshot storage, per GB-month of the snapshot's size, while the snapshot exists, when the region has a price for it. |
A snapshot that fails before it becomes available isn't billed. Every resource is billed for at least 60 seconds over its lifetime, once. For how usage becomes charges on your invoice, see How billing works. For prices, see Pricing.
Limits
- A restore point copies every disk of the instance. You can't select single disks.
- A volume snapshot is the full size of its source volume. Linking it into a chain doesn't make it smaller.
- You can take a volume snapshot of a volume that's Available or In use.
- Volume snapshots count toward the subscription's quota and the project's allocation. See Block Storage limits.
Related services
- Virtual Machines: instances that restore points copy and that you restore to.
- Block Storage: volumes that volume snapshots copy and that you restore to.
- Images: reusable images for launching new instances.
- Backup Center: scheduled backups with a retention period.