Restoring a Backup to a Specific Point in Time

You can restore an existing automated backup to a specific point in time.

The most recent full backup will be downloaded from OBS for restoration. After the restoration is complete, incremental backups will be replayed to the specified point in time. The time required depends on the amount of data to be restored.

Precautions

GeminiDB Cassandra instances allow you to restore data to a new instance at a specific point in time.
After automated backup is enabled, the system performs an incremental backup based on the preset incremental backup interval. The incremental backup is stored in OBS.

Data can be restored to a specified time point only after the automated backup policy is enabled.
During the instance restoration, backup files are downloaded from OBS buckets to the data directory of the restored instance. The download bandwidth of OBS is 40 MB/s.

Procedure

Log in to the GeminiDB console.
In the service list, choose Databases > GeminiDB.
On the Instances page, click the instance that you want to perform a PITR for.
In the navigation pane on the left, choose Backups & Restorations.
On the Backups & Restorations page, click Restore to Point in Time.

Figure 1 Restoring data to a point in time
Select a restoration date and a time point to which data is restored, and then click Yes.

Figure 2 Restore to Point in Time
On the Create New Instance page, create an instance of the same specifications as the instance to be restored. The new instance is independent from the original one.
- The new instance should be deployed in a different AZ to ensure that your applications will not be affected by SPOFs.
- The compatible API, instance type, instance version, and CPU type are the same as those of the original and cannot be changed.
- Other settings are the same as those of the original instance by default but can be modified. For details, see Buying an Instance.