如何使用计算巢备份-Compute Nest(Compute Nest)-阿里云帮助中心

Background

You may need to back up data for several reasons. For example:

For a dedicated Palworld server, game progress can be lost due to issues within the game. Regular backups are necessary to prevent this.
For a database service, you need to create backups to ensure data security and prevent data loss from potential failures.
Before upgrading a service, you must back up your data to prevent data loss during the upgrade.

Previously, Compute Nest users had to manually back up service instance data in the console of each cloud product. For example, if a service instance was deployed on an ECS instance, you had to log on to the ECS console, filter for the specific ECS instance, and manually create a snapshot for its disk. To restore, you had to find the correct snapshot and perform a data rollback in the ECS console. This process becomes complex and error-prone when managing numerous disks or ApsaraDB RDS instances. To solve this, Compute Nest introduced its backup management feature. This feature allows you to back up or restore cloud resource data for your service instances with a single click directly from the Compute Nest console.

How it works

Support for a specific cloud product is available only if Compute Nest is integrated with it and the product itself supports these operations. Currently, Compute Nest supports backup and restore for the following cloud products:

Elastic Block Storage (EBS).
ApsaraDB RDS instances that run MySQL, SQL Server, or PostgreSQL.

Backup

Before a data backup begins, Compute Nest creates an empty backup record. This record stores the start time, end time, status, and the instance IDs and backup status of all cloud products involved in the backup.

When you initiate a data backup, Compute Nest uses the service instance ID and resource tags to identify all associated cloud resources. It then filters for Elastic Block Storage (EBS) disks and the specified types of ApsaraDB RDS instances. Next, Compute Nest calls the APIs of the respective cloud products to create the backups. After the API calls succeed, Compute Nest polls the backup status from each cloud product. When the status of a backup task becomes final (success or failure), the result is recorded and displayed. A single Compute Nest backup task typically involves multiple instances across several cloud products. As long as any instance is in the "Backing Up" state, the overall Compute Nest backup record remains in the "Backing Up" state. After all instance backups are complete, the overall backup record is marked as "Backup Failed" if any individual backup fails. The record is marked as "Backup Successful" only when all individual backups succeed.

Restore

Similar to the backup process, Compute Nest creates an empty restore record before starting. This record stores the start time, end time, the ID of the backup being used, the overall restore status, and the resource IDs and restore status for each cloud product involved.

To restore data from an EBS disk, the attached ECS instance must be stopped. Therefore, Compute Nest automatically stops the entire service instance before starting the restore. Once the instance is stopped, Compute Nest identifies the backup to be used and calls the APIs of the respective cloud products to restore the data. After the API calls succeed, Compute Nest polls each cloud product for the restore status. When a task reaches a final state (success or failure), the result is recorded and displayed. A single restore task often involves multiple instances across multiple cloud products. As long as any instance is in the "Restoring" state, the overall restore record remains in the "Restoring" state. When all instances are restored, the overall record is marked as "Restore Failed" if any individual restore fails. The record is marked as "Restore Successful" only when all individual restores succeed. After the restore is complete, Compute Nest automatically restarts the service instance. A specific mechanism for restoring ApsaraDB RDS instances that run MySQL or PostgreSQL may rename databases and tables, making them temporarily inaccessible. To resolve this, Compute Nest uses Function Compute (FC) to execute SQL statements that restore the original database and table names.

Important

If your service instance includes ApsaraDB RDS instances that run MySQL or PostgreSQL, you may incur Function Compute (FC) fees during the restore process.

Billing

The Compute Nest backup feature is free. However, you are billed by the underlying cloud products for the resources they use. Costs typically come from three sources:

Snapshot fees for Elastic Block Storage (EBS). For more information, see Snapshot billing.
Fees for ApsaraDB RDS backups. For more information, see:
- MySQL engine: Backup fees.
- SQL Server engine: Backup fees.
- PostgreSQL engine: Backup fees.
Function Compute (FC) fees. For more information about why these fees may be generated, see the How it works > Restore section. For details on FC billing, see Billing overview.

Procedure

If the Backup and Restore feature is enabled for your service, you can manage backups and restores from the service instance details page.

Backup

Go to the Service instance page and click a service instance ID to view its details.
Go to the Backup and Restoration tab and click Create Backup. Click the Backup and Restoration tab, and on the Backup Tasks subtab, click Create Backup.

Warning
During the backup, do not modify the service instance or its data.
After the backup is created, you can view its details, such as the backup ID, type, start/end time, and status in the Backup Task list. For completed backups, you can click Restore or Delete. If a backup fails, you can hover over its status to view the reason.

Restore

In the Actions column for the desired backup, click Restore.
Click OK to start the restore. You can view the list of restore tasks and their status on the Restoration Task tab. The initial status of a task is Waiting.
Warning
- A restore operation stops the service instance. We recommend creating a new backup before you begin.
- During the restore, do not modify the service instance or its data.
Wait for the restore to complete. If it fails, the status is displayed as Restore Failed. You can hover over the status to view the error details. The restore task list includes columns such as Restore ID, Start Time, End Time, Status, Creation Time, and Update Time.