Restore a released cluster

Usage notes

• Each cluster in the Cluster Recycle Bin must have at least one backup set. If you delete all backup sets for a cluster, the cluster cannot be restored.

• After you release a cluster, it is asynchronously archived as a level-2 backup in the Cluster Recycle Bin at a rate of about 150 MB/s. For more information about backups, see backup and restoration.

Procedure

1. Log on to the PolarDB console.

2. In the left-side navigation pane, click Cluster Recycle Bin.

3. Select the region where the cluster is located. Find the target cluster and click Restore Data to New Cluster in the Actions column.

4. On the Custom Purchase page, configure the following parameters:

   Parameter	Description
   Billing Method	Subscription: A prepaid billing method. When you create a cluster, you select and pay for resources of a fixed specification in advance. Longer subscription periods provide larger discounts. This billing method is suitable for scenarios in which workloads are stable over time. Pay-as-you-go: A postpaid billing method. When you create a cluster, you select resources of a fixed specification but do not pay in advance. You are billed based on the actual usage duration. This billing method is suitable for scenarios in which workloads change frequently. Note If the current billing method of your PolarDB cluster does not meet your business requirements, you can change the billing method at any time.
   Region	Select the geographical location of the deleted cluster. The region cannot be changed after purchase. Note Ensure the PolarDB cluster and the ECS instance you connect from are in the same region. Otherwise, they must communicate over the public internet, which degrades performance. You can create PolarDB and ECS in the same availability zone or in different availability zones. Creating PolarDB and ECS in the same availability zone can reduce network latency and improve access speed.
   Creation Method	Select Restore from Recycle Bin. This option restores a deleted cluster from the Cluster Recycle Bin.
   Original Version	Select the version of the deleted cluster.
   Deleted Clusters	Select the name of the deleted cluster.
   Historical backups	Select the backup that you want to restore from. Note The Historical Backups list shows times in UTC, while the main backup list uses your local system time. Ensure you select the correct backup. For example, if the backup list shows 10:00:00 on May 8, 2020 (UTC+8), the corresponding time in Historical Backups is 2020-05-08T02:00:00Z.
   Database Engine	The database engine version of the PolarDB cluster. It defaults to the version of the deleted cluster and cannot be changed.
   Service Edition	The product edition of the PolarDB cluster. Select Standard Edition or Enterprise Edition.
   Specification Edition	Two specifications are supported: General-purpose and Dedicated. Dedicated: Each cluster exclusively uses its allocated compute resources, such as CPU, and does not share resources with other clusters on the same server. This provides more stable and reliable performance. General-purpose: Different clusters on the same server share idle compute resources, such as CPU. This improves cost-efficiency by reusing compute resources.
   CPU Architecture	Set CPU Architecture to X86 or YiTian ARM. X86: The X86 architecture uses Intel processors and a high-performance network. This improves overall performance and stability and meets enterprise requirements for high stability and compute performance. ARM: The ARM architecture uses Alibaba Cloud's self-developed YiTian 710 processor and a 25 GE intelligent high-speed NIC to deliver powerful computing capabilities. Combined with a high-performance network, it meets the demands for cost-effectiveness, security, and stability for cloud workloads in sectors like government and the internet industry.
   Primary Zone	Select a Primary Zone for the cluster. An availability zone is an independent physical area within a region. There are no significant differences between availability zones. You can create the PolarDB cluster and the ECS instance in the same or different availability zones.
   Network Type	Configure the VPC and vSwitch. The network type is fixed to VPC. Make sure that the PolarDB cluster and the ECS instance that you want to connect from are created in the same VPC. Otherwise, they cannot communicate over an internal network, which can degrade performance. If you already have a VPC that meets your network requirements, select it. For example, if you have an existing ECS instance and its VPC meets your requirements, select that VPC. If you do not have a VPC that meets your network requirements, you can use the default VPC and vSwitch: Default VPC: It is unique in the selected region. It uses a 16-bit subnet mask (e.g., 192.168.0.0/16), providing up to 65,536 private IP addresses. It does not count against the VPC quota allocated by Alibaba Cloud. Default vSwitch: It is unique in the selected availability zone. It uses a 20-bit subnet mask (e.g., 192.168.0.0/20), providing up to 4,096 private IP addresses. It does not count against the vSwitch quota in your VPC. If the default VPC and vSwitch do not meet your requirements, you can create a VPC and vSwitch. For more information, see Create and manage a VPC.
   High availability mode	Select a HA Mode. Dual AZ (Enable Hot Standby Storage Cluster): PolarDB deploys both a primary cluster (including primary/standby compute nodes and primary storage) and a hot standby storage cluster in the same region. Each cluster maintains three data replicas (for a total of six replicas), ensuring a higher SLA. Single AZ (Disable Hot Standby Storage Cluster): Only the primary cluster (with three data replicas) is deployed. Storage costs are 50% lower than in the dual-AZ setup, but the cluster's SLA is also lower. Note For more information about the hot standby storage cluster and its solutions, see High availability mode.
   Secondary Zone	Select a Secondary Zone. This parameter is required only when HA Mode is set to Dual AZ (Enable Hot Standby Storage Cluster). You can select Automatically Allocated.
   Nodes	By default, the system provisions two nodes: one primary and one read-only. You can keep this configuration or select only a single primary node.
   PolarProxy type	Two PolarProxy types are supported: General-purpose Enterprise Edition and Dedicated Enterprise Edition. The available type depends on the cluster specification you choose. General-purpose Enterprise Edition: Paired with the general-purpose specification. It shares physical CPU resources and provides intelligent, second-level resource elasticity based on workload. Dedicated Enterprise Edition: Paired with the dedicated specification. It exclusively occupies physical CPU resources for better performance stability.
   Storage Type	PolarDB supports the following storage types: ESSD, PSL5, and PSL4. ESSD: Alibaba Cloud’s ultra-high performance disk product. Built on a next-generation distributed block storage architecture, it uses a 25GE network and RDMA technology to deliver up to 1 million random IOPS per disk and reduced single-path latency. ESSDs are available in the following performance levels: PL0 ESSD: An ESSD at performance level 0 (PL0). ESSD PL1: Compared with PL0, PL1 increases IOPS by up to 5 times and throughput by approximately 2 times. ESSD PL2: Compared with PL1, PL2 increases IOPS and throughput by approximately 2 times. ESSD PL3: Compared with PL2, PL3 increases IOPS by up to 10 times and throughput by up to 5 times. It is suitable for workloads that require extremely high concurrent I/O performance and highly stable read/write latency. ESSD AutoPL: Compared with other PLs, AutoPL decouples IOPS from capacity. You can flexibly configure and adjust performance on demand to reduce the total cost of ownership (TCO). PSL5: A storage type supported in previous versions of PolarDB. It provides high performance, reliability, and availability. PSL4: A new storage type for PolarDB that uses Alibaba Cloud's proprietary smart-SSD technology. It compresses and decompresses data at the physical SSD level, reducing the storage price per unit of data with a controllable impact on performance. Note When an ESSD runs out of storage capacity, the disk becomes read-only. For a detailed comparison of PSL5 and PSL4, see How to choose between PSL4 and PSL5. The storage type of an existing cluster cannot be changed. To switch types, you must purchase a new cluster with the desired storage type and migrate your data.
   Storage Billing Method	PolarDB supports two storage billing methods: Pay-by-Capacity (Pay-as-you-go) and Subscription. Pay-by-Capacity (Pay-as-you-go): A serverless method where you do not need to specify a storage capacity at purchase. Storage automatically expands as your data grows, and you are charged only for the storage space that your data occupies. Subscription: This method requires you to prepay for a specific storage capacity when you purchase the cluster. Note This parameter can be configured only when the cluster Billing Method is Subscription and the Service Edition is Enterprise Edition. When the cluster Billing Method is Subscription and the Service Edition is Standard Edition, this parameter is fixed to Subscription.
   Storage Capacity	The amount of storage capacity to pre-purchase. The storage capacity range depends on the compute node specification and storage type: For PSL4/PSL5, the range is 10 GB to 500 TB, with a minimum increment of 10 GB. For ESSD, the range is 20 GB to 64 TB, with a minimum increment of 10 GB. Note This parameter is available only when the cluster Billing Method is Subscription and the Storage Billing Method is Subscription.

5. On the right side of the page, confirm that all configuration items, such as Billing Method, Region, Deployment Topology, and Resource Group, meet your business requirements. Then, configure the following parameters:

   Parameter	Description
   Quantity	Select the number of clusters to purchase. Note You can create up to 50 clusters in a single order, which is suitable for scenarios such as batch server launches for games. Each Alibaba Cloud primary account can purchase a maximum of 50 clusters.
   Subscription Duration	Configure the Subscription Duration for the cluster and select whether to Enable Auto-renewal. This is required only when the billing method is Subscription. Note We recommend selecting Enable Auto-renewal to prevent service interruptions from a missed payment.

   Review the total cost and the cost breakdown. By clicking Buy Now, you agree to the relevant terms of service. On the Pay page, confirm the order and click Pay to complete the purchase.

   The cluster is created within 10 to 15 minutes of a successful purchase. You can then find the new cluster in the cluster list.

   Note

   • The cluster is unavailable while its nodes are in the Creating state and becomes ready for use only when its status changes to Running.

   • Ensure you have selected the correct region; otherwise, you will not find the newly created cluster.

API reference