Snapshot overview

更新时间:
复制 MD 格式

A snapshot is a point-in-time backup of disk data. The first snapshot for a disk is a full snapshot that captures all written data blocks (empty blocks are excluded). Subsequent snapshots of the same disk are incremental snapshots that capture only the data blocks changed since the previous snapshot.

ENS snapshots are crash-consistent: only data already written to disk at the moment of creation is captured. Data held in memory or pending I/O operations is not included.

Important

ENS does not automatically create snapshots. All snapshots must be created manually. Create snapshots before high-risk operations such as configuration changes, upgrades, or migrations so you can recover if something goes wrong.

Features

image

Feature

Description

Create a snapshot

You can create snapshots in the following ways:

  • Manual snapshot: Create a snapshot for a single cloud disk on demand. Suitable for one-time checkpoint backups, such as before system upgrades or configuration changes.

  • Auto snapshot: Create an auto snapshot policy and apply it to a cloud disk. The system automatically creates snapshots for the disk according to the policy's schedule and retention settings. For details, see Auto snapshot policy.

Roll back a disk

Restore a disk to a previous state using a snapshot when the system is unresponsive or accidental changes are made.

Create an image from a system disk snapshot

Create a custom image from a snapshot that contains the OS and data of an ENS instance, then use that image to create multiple instances with identical environments.

Create a disk from a snapshot

Create multiple identical disks from a single snapshot.

Replicate a snapshot across nodes

Copy a snapshot from one node to another to create a replica in the destination node.

Limitations

  • Snapshots can only be created for ultra disks or all-flash disks.

  • Images cannot be created from data disk snapshots.

  • Snapshots can only be replicated to nodes that support ultra disks or all-flash disks.

  • A single account can retain a maximum of 100 snapshot policies.

  • Each cloud disk can be associated with a maximum of 1 automatic snapshot policy.

  • Each cloud disk can retain a maximum of 100 automatic snapshots.

Billing

Snapshot storage is billed on a pay-as-you-go basis by month. You are charged based on the actual storage space the snapshot occupies — not the provisioned disk size.

Storage usage

Formula: Storage fee = Unit price × Storage usage (GB) × Billing duration (months)

Disk type

Unit price

Ultra disk

USD 0.0493 per GB-month

All-flash disk

USD 0.1408 per GB-month

Billing duration: Starts when the snapshot is created and ends when it is deleted. A snapshot that exists for 21 days counts as 21 ÷ 30 = 0.7 months.

Cross-node snapshot replication

When you replicate a snapshot to another node, charges apply to both the replication traffic and the snapshot storage in the destination node.

  • The first replication of a disk snapshot is a full copy containing all data on the disk.

  • Subsequent replications of the same disk to the same node are incremental, capturing only data changed since the previous replica.

Formula: Replication traffic fee = USD 0.05 per GB × Storage usage (GB) × 2

The multiplier of 2 accounts for traffic generated in both the source node and the destination node.

Billing examples

Example of snapshot size billing

On the first day of a month, User B creates a total of 100 GB of snapshots for an ultra disk and releases all of them on the 21st day. Also on the first day, User B creates 50 GB of snapshots for an all-flash cloud disk and keeps them for the entire month.

Therefore, the snapshot capacity fee for the month is CNY 0.35/GB/month × 100 GB × (21/30) + CNY 1/GB/month × 50 GB × 1 month = CNY 74.5.

Example of cross-edge-zone snapshot replication traffic billing

On a given day in a calendar month, User B copies a 100 GB snapshot from an edge zone in Hangzhou to an edge zone in Beijing. This operation creates a 100 GB snapshot replica in the Beijing edge zone.

The cross-edge-zone snapshot replication traffic fee is calculated as follows: USD 0.05/GB × 100 GB × 2 = USD 10.

Use cases

  • Daily data backup: Use an auto snapshot policy to periodically create snapshots for your cloud disks and retain them for a specified duration. This provides routine data backup to protect against data loss caused by accidental operations, virus infections, or malicious attacks.

  • Protection for high-risk operations: Before performing scaling, specification changes, or migrations on your ENS instances and cloud disks, create a manual snapshot in advance. If issues arise during the operation, you can quickly recover your business from the snapshot.

  • Quick data recovery: If your system experiences service interruptions due to software compatibility issues or other causes during upgrades, you can roll back the cloud disk from a historical snapshot to quickly restore business continuity. For details, see Create and manage snapshots.

  • Batch deployment across sites: Create a custom image from a snapshot to efficiently deploy or upgrade multiple sites with identical environments in bulk.

  • Batch data distribution: Create new cloud disks in bulk from snapshots and attach them to ENS instances as needed for rapid data distribution across your infrastructure.

What's next

  • Create a snapshot

  • Roll back a disk using a snapshot

  • Replicate a snapshot across nodes