Back Up and Restore TiDB Cluster Data

This document describes how to back up and restore your TiDB cluster data on TiDB Cloud.

Note

For Serverless Tier clusters, the backup and restore feature is unavailable. You can use Dumpling to export your data as a backup.

Currently, the backup and restore feature has the following restrictions:

TiDB Cloud does not support restoring tables in the mysql schema, including user permissions and system variables. You can use Dumpling and Lightning to manually back up and restore these data.
If you turn on and off PITR (Point-in-time Recovery) multiple times, you can only choose a time point within the recoverable range after the most recent PITR is enabled. The earlier recoverable range is not accessible.
You cannot disable automatic backup.

Backup

TiDB Cloud supports automatic backup and manual backup.

Automatic backups are scheduled for your TiDB clusters according to the backup setting, which can reduce your loss in extreme disaster situations. You can also pick a backup snapshot and restore it into a new TiDB cluster at any time.

Automatic backup

By the automatic backup, you can back up the cluster data every day at the backup time you have set. To set the backup time, perform the following steps:

Navigate to the Backup tab of a cluster.
Click Backup Settings. The setting window displays.
In the setting window, configure the automatic backup:
- Turn on or off the PITR (Point-in-time Recovery) feature.
  PITR supports restoring data of any point in time to a new cluster. You can use it to:
  - Reduce RPO in disaster recovery.
  - Resolve cases of data write errors by restoring point-in-time that is before the error event.
  - Audit the historical data of the business.
    If you have one of the preceding needs and want to use the PITR feature, make sure that your TiDB cluster version is at least v6.4.0 and the TiKV node size is at least 8 vCPU and 16 GiB.
- In Backup Time, schedule a start time for the daily cluster backup.
  It is recommended to schedule automatic backup at a low workload period. If you do not specify a preferred backup time, TiDB Cloud assigns a default backup time, which is 2:00 AM in the time zone of the region where the cluster is located.
- In Backup Retention, configure the minimum backup data retention period.
- In Backup Storage Region, select the regions where you want to store your backup data.
  TiDB Cloud stores your backup data in the current region of your cluster by default. This behavior cannot be changed. In addition, you can add another remote region, and TiDB Cloud will copy all new backup data to the remote region, which facilitates data safety and faster recovery. After adding a remote region as a backup data storage, you cannot remove the region.
Click Confirm to preview the configuration change.
If you turn on PITR, you can select the Perform a backup immediately and use it as recovery starting point in PITR. checkbox. Otherwise, PITR will not be available until the next backup is completed.
Click Confirm.

Backup storage region support

Currently, you cannot select an arbitrary remote region for backup data storage. The regions already supported are as follows:

Cloud provider	Custer region	Remote region support
GCP	Tokyo (asia-northeast1)	Osaka (asia-northeast2)

Note

If you select multiple backup storage regions, you will be charged for multiple backup storage and inter-region backup data replication out from the cluster region to each destination region. The cost is on a per-region basis and varies with the backup regions selected. For more information, see Data Backup Cost.

Manual backup

Manual backups are user-initiated backups that enable you to back up your data to a known state as needed, and then restore to that state at any time.

To apply a manual backup to your TiDB cluster, perform the following steps:

Navigate to the Backup tab of a cluster.
Click Manual Backup. The setting window displays.
Enter a Name.
Click Confirm. Then your cluster data is backed up.

Delete backup files

To delete an existing backup file, perform the following steps:

Navigate to the Backup tab of a cluster.
Click Delete for the backup file that you want to delete.

Delete a running backup job

To delete a running backup job, it is similar as Delete backup files.

Navigate to the Backup tab of a cluster.
Click Delete for the backup file that is in the Pending or Running state.

Best practices for backup

It is recommended that you perform backup operations at cluster idle time to minimize the impact on business.
Do not run the manual backup while importing data, or during cluster scaling.
After you delete a cluster, the existing manual backup files will be retained until you manually delete them, or your account is closed. Automatic backup files will be retained for 7 days from the date of cluster deletion. You need to delete the backup files accordingly.

Restore

TiDB Cloud provides two types of data restoration:

Restore backup data to a new cluster
Restore a deleted cluster from the recycle bin

Restore data to a new cluster

To restore your TiDB cluster data from a backup to a new cluster, take the following steps:

Navigate to the Backup tab of a cluster.
Click Restore. The setting window displays.
In Restore Mode, you can choose to restore data of any point in time or a selected backup to a new cluster.
- Select Time Point
- Select Backup Name
To restore data of any point in time within the backup retention to a new cluster, make sure that PITR in Backup Settings is on and then take the following steps:
1. Click Select Time Point.
2. Select Date and Time you want to restore to.
To restore a selected backup to the new cluster, take the following steps:
1. Click Select Backup Name.
2. Select a backup you want to restore to.
In Restore to Region, select the same region as the Backup Storage Region configured in the Backup Settings.
In the Restore window, you can also make the following changes if necessary:
- Set the cluster name.
- Update the port number of the cluster.
- Increase the node size, node quantity, and node storage for the cluster.
Click Restore.
The cluster restore process starts and the Security Settings dialog box is displayed.
In the Security Settings dialog box, set the root password and allowed IP addresses to connect to your cluster, and then click Apply.

Restore a deleted cluster

To restore a deleted cluster from recycle bin, take the following steps:

Log in to the TiDB Cloud console.
In the left navigation pane of the Clusters page, do one of the following:
- If you have multiple projects, switch to the target project, and then click Admin > Recycle Bin.
- If you only have one project, click Admin > Recycle Bin.
Locate the cluster you want to restore, and then click Backups in the Action column.
Locate your desired backup time, and then click Restore in the Action column.
In the Restore window, make the following changes if necessary:
- Update the port number of the cluster.
- Increase the node size, node quantity, and node storage for the cluster.
Click Confirm.
The cluster restore process starts and the Security Settings dialog box is displayed.
In the Security Settings dialog box, set the root password and allowed IP addresses to connect to your cluster, and then click Apply.

Note

You cannot restore a deleted cluster to any point in time. You can only select an automatic or manual backup to restore.