A cloud disk has adopted a three-copy redundancy mechanism for data security. Why do we still need to use snapshots?

In situations where a logic-level data error occurs, for example, suppose a user deletes data by mistake, or if the data is damaged by a virus or file system exceptions, all three copies of the data stored will be affected and historical data cannot be restored. If you have created a snapshot previously, you can use the snapshot to restore data to the point in time the snapshot was created.

Suppose an administrator created snapshot A for a cloud disk at 11:00, and the cloud disk was infected with a virus at 12:00, which caused that data to be unusable. In this case, the three copies of the data would have been updated to state 2, and the data cannot be restored. Only snapshot A created at 11:00 can be used to restore data to the uninfected state 1.

What are the differences between a snapshot and an image? What is the relationship between the two?

Differences between a snapshot and an image

If no data disk is mounted to an instance and all data is written on the system disk, the data on the system disk cannot be protected by creating an image. Images cannot be scheduled for continuous backup. Once the system disk data is damaged, you can only recover the data to when the image was initially created. Therefore, images are not suitable for data protection. Specific differences are as follows:

Relationship between a snapshot and an image

• The CBS snapshot service is used for storing underlying data of an image. The associated snapshot generated by a custom image is billed based on the actual capacity of the snapshot. You can view the total snapshot capacity in the snapshot overview.
• The image and its associated snapshot has a binding relationship. You must first delete the image before deleting its associated snapshot. You can check image details for information on the snapshot that is associated with the custom image.
• You can create snapshots for both the system disk and a data disk. However, you can only create a custom image from the snapshot of the system disk. The custom image can then be used to create one or more instances with the same environment.
• When using shared images, only the Tencent Cloud account that created the shared image needs to pay for associated snapshot fees. The shared user does not need to pay.

What is a scheduled snapshot?

A scheduled snapshot is automatically created by the system according to the scheduled snapshot policy you created for the cloud disk associated with the policy. To use this feature, you must first create a scheduled snapshot policy and associate it with the cloud disk.

What limits do scheduled snapshots have?

A maximum of 30 scheduled snapshot policies can be created in one region. Each scheduled snapshot policy can be associated with up to 200 cloud disks. In addition, snapshots created according to the scheduled snapshot policies comply with snapshot quota limits. For more information, see Use Limits.

Is the snapshot feature available in all availability zones?

The snapshot feature is available in all availability zones.

How are snapshots billed?

Snapshots are billed according to your total snapshot storage capacity in each region. CBS is pay-as-you-go with an hourly billing cycle. For more information on billing, see Billing Overview and Pricing List.

Do I need to unmount a disk or interrupt all reads and writes before creating a snapshot?

No. You can create a real-time snapshot while the disk is connected and in use, without affecting normal business operation. However, the snapshot can only capture data that has been written into the cloud disk, not data cached in memory by applications or the operating system. To ensure all application data is captured, we recommend that you suspend all disk I/O operations before creating a snapshot. For a cloud disk that is used as a system disk, we recommend that you shut down the CVM to create a more complete snapshot.

Does snapshot creation affect disk performance?

Creating a snapshot will occupy a small amount of the disk I/O. We recommend creating snapshots when business is slow.

How long does it take to create a snapshot?

The time for creating a snapshot is subject to factors such as the number of disk writes, and the underlying read-write operations. It is hard to predict, but creating a snapshot does not affect your normal disk use.

How do I create a cloud disk by using a snapshot?

See Creating Cloud Disks using Snapshots.

How do I roll back snapshots?

See Rolling Back Snapshots.

Do I need to shut down the CVM to roll back a snapshot?

• For a cloud disk that has been mounted to a CVM, you have to shut down the CVM during rollback.
• For a cloud disk that is not mounted, you can directly perform rollback.

Can I read a previous snapshot to restore a cloud disk?

Yes. You can use an existing snapshot created at any point in time to restore data, regardless of the snapshot’s point in time.

Can I delete the source snapshot when it is being replicated?

No. It can only be deleted after the replication is complete.

Is the new snapshot still associated with the source snapshot’s source disk after replication is complete?

After cross-region replication, the new snapshot is no longer associated with the source disk of the source snapshot. The rollback feature is unavailable for replicated snapshot.

How do I delete a snapshot?

• Cloud disk snapshots can be deleted on the console or via APIs. For more information, see Deleting Snapshots.
• For snapshots associated with custom images, you must delete the custom image before deleting the snapshot.