PolarDB:Restoration method 1: Restore data to a specific point in time

Restore a cluster to a point in time

1. Log on to the PolarDB console. In the left-side navigation pane, click Clusters. Select a region in the upper-left corner, then click the cluster ID to open the Basic Information page.

2. In the left-side navigation pane, choose Settings and Management > Backup and Restoration.

3. Start a point-in-time restore based on your restore scope:

   • Within a region: On the Backup and Restoration page, click Point-in-time Restoration.

   • Across regions: On the Backup and Restoration page, select the region where the backup data resides, then click Point-in-time Restoration.

4. On the Clone Instance page, select a billing method for the new cluster.

   Billing method	Description
   Subscription	Pay for compute nodes upfront. Storage fees are deducted from your account balance hourly.
   Pay-as-you-go	Pay for compute nodes and storage hourly after use. Suitable for short-term recovery work where you plan to release the cluster after migrating data back.

5. Configure the following parameters.

   Parameter	Description
   Clone type	Select Restore to Point in Time.
   Region	The region to restore data into. Cross-region backup supports restoring to either the source or destination region. If cross-region backup is not enabled, the region matches the original cluster and cannot be changed.
   Point in time	The timestamp to restore to. PolarDB supports point-in-time restore for any moment in the past 7 days.
   Primary zone	The zone where the new cluster's primary node runs. In regions with two or more zones, PolarDB automatically replicates data to a secondary zone for disaster recovery.
   Network type	Fixed to VPC. No configuration required.
   VPC	Select a virtual private cloud (VPC) and a vSwitch for the new cluster. Use the same VPC and vSwitch as your original cluster when possible. Make sure the PolarDB cluster and the Elastic Compute Service (ECS) instance you plan to connect are in the same VPC—otherwise they cannot communicate over the internal network.
   vSwitch
   Compatibility	Automatically set from the original cluster. No change needed.
   Product edition	Automatically set from the original cluster. No change needed.
   Edition	Automatically set from the original cluster. No change needed.
   Specification type	Automatically set from the original cluster. No change needed.
   CPU architecture	Automatically set from the original cluster. No change needed.
   Node specification	The compute capacity for each node. We recommend that you select a node specification that is higher than the node specification of the original cluster to ensure stable performance. For available options, see Specifications of compute nodes.
   Nodes	Default is 2 (one primary node and one read-only node). After creation, you can add up to 15 read-only nodes. See Add or remove a read-only node.
   PolarProxy type	Fixed to Standard Enterprise. No change needed.
   Enable hot standby cluster	When enabled, both the primary cluster and a hot standby storage cluster are deployed in the same region, each with 3 replicas (6 total). This delivers a higher service-level agreement (SLA). When disabled, only the primary cluster is deployed with 3 replicas, and the storage unit price is half the cost of the enabled option.
   Storage type	PolarDB Enterprise Edition supports the PSL5 and PSL4 storage types. PSL5 is the established storage type with high performance, reliability, and availability. PSL4 uses Smart-SSD technology to compress data at the physical SSD level, reducing storage costs while maintaining disk performance. Storage type cannot be changed after cluster creation. For a detailed comparison, see Comparison between PSL4 and PSL5.
   Storage billing method	Pay-as-you-go: Storage scales automatically with data volume. No upfront capacity purchase required. Subscription: Purchase a fixed storage capacity upfront. Available only when the billing method is Subscription. If the billing method is Pay-as-you-go, the Storage Billing Method parameter is unavailable and storage is always charged on a pay-as-you-go basis. For more information, see Billing methods for storage.
   Storage cost	Applies when storage billing method is Pay-as-you-go. You are charged for actual storage usage hourly. To reduce costs, purchase a storage plan. See Purchase a storage plan.
   Storage capacity	Applies when both billing method and storage billing method are Subscription. Enterprise Edition supports 50 GB to 500 TB in 10 GB increments.
   Cluster name	2–128 characters. Must start with a letter and can contain letters, digits, periods (.), underscores (_), and hyphens (-). Leave blank to auto-generate a name. You can change the name after creation.
   Purchase plan	The subscription duration. Applies only when the billing method is Subscription.
   Number	The number of clusters to create.

6. Complete the purchase:

   • Pay-as-you-go: Accept the Terms of Service, then click Buy Now.

   • Subscription: Accept the Terms of Service, click Buy Now, then confirm the order and payment method on the Purchase page and click Purchase.

   Note

   Cluster creation takes 10 to 15 minutes. After that, the new cluster appears on the Clusters page.

How long does restore take?

Restore time is the sum of two phases:

1. Snapshot restore: PolarDB restores data from the most recent backup set (snapshot) before your target point in time.

2. Physical log replay: PolarDB replays physical logs from the snapshot to your exact target timestamp.

If the snapshot before your target time is older and write activity was high during that interval, the log replay phase takes longer. To understand the log backup mechanism, see Physical log backup.

API reference