When your workload has cyclical fluctuations or requires planned adjustments, you can manually upgrade or downgrade the compute nodes (CPU and memory) of your PolarDB cluster. This helps you meet performance demands while controlling costs.
Impact assessment
Before changing specifications, assess the potential impact on your business.
Service impact
Changing the specifications of a PolarDB cluster triggers a node restart, which causes transient connections and brief performance fluctuations. The impact depends on whether the cluster supports fast failover (hot standby) and whether Binlog is enabled.
-
If the cluster does not support fast failover (hot standby), a transient connection lasting about 20 to 30 seconds may occur during the specification change. Therefore, ensure that your application has a connection retry mechanism in place before the switchover.
-
If the cluster supports fast failover (hot standby), a specification change with Binlog enabled may cause a transient connection lasting about 5 to 10 seconds. If Binlog is disabled, performance may drop to zero for 1 to 3 seconds, but no transient connections occur and transactions are not interrupted.
-
Change duration: The change task takes about 5 minutes per compute node. For example, changing two nodes takes about 10 minutes. The actual duration depends on factors such as cluster workload and the number of databases and tables.
-
Data latency: During the change, replication latency on read-only nodes may be longer than usual.
-
Data security: The specification change only adjusts compute node resources and does not affect any existing data in the cluster.
-
Restart sequence: When you modify the specifications of both the primary node and read-only nodes at the same time, the system restarts the read-only nodes first, and then restarts the primary node. No primary/secondary switchover occurs during this process.
Node specification limitations
To ensure cluster stability and high availability, the following limitations apply to the specifications of the primary and read-only nodes:
-
Make sure that at least one read-only node has the same specifications as the primary node.
-
Read-only nodes with hot standby enabled must have the same specifications as the primary node.
-
A read-only node must have at least half the memory of the primary node.
-
Read-only nodes have minimum CPU core requirements that correspond to the CPU specifications of the primary node:
Primary node CPU cores
Minimum read-only node CPU cores
2
2
4
4
8
4
16
8
32
16
64
32
88
64
120
64
-
The Multi-master Cluster (Limitless) edition is not subject to the preceding limitations.
-
Due to the analytical workloads they handle, a columnar index read-only node should have higher specifications than the primary node.
Prerequisites
Before you begin, ensure the following prerequisites are met:
-
Billing method: The target cluster must use the Subscription or Pay-as-you-go billing method.
-
Cluster status: The target cluster must be in the Running state, and no other configuration change tasks, such as adding or deleting nodes, or minor version upgrades, are in progress.
Procedure
-
Log on to the PolarDB console. In the left-side navigation pane, click Clusters. Select the region where your cluster is located.
-
Navigate to the Change Configurations page by using one of the following methods:
-
In the Actions column for the target cluster, click Change Configurations.
-
Click the target cluster ID to go to the Basic Information page. In the Database Nodes section, click Change Configurations.
-
-
In the Change Configurations dialog box, select Upgrade or Downgrade, and then click OK.
-
On the Upgrade/Downgrade page, confirm the Current Configuration and set the following parameters:
Parameter
Description
Edition
Select the target specification edition. Options include General Specification and Dedicated Specification.
Node
Select new specifications for the nodes you want to change. You can apply the same specifications to the entire cluster or configure each node individually.
NoteMake sure that one read-only node has the same specifications as the primary node. You can configure the specifications of other nodes as needed.
Switchover time
Select when you want the configuration change to take effect.
-
Effective Immediately: The change takes effect immediately.
-
Scheduled Switchover: Select a time within the next 24 hours. The change is completed within 30 minutes after the specified time. You can view or cancel the task on the Scheduled Tasks page.
-
-
Read the service agreements, and then click Buy Now to complete the payment. After payment is complete, the system executes the task at the specified switchover time.
Billing
Specification changes affect your costs. The billing rules vary based on the billing method of the cluster.
-
Pay-as-you-go: After a successful specification change, you are billed hourly for the new specifications.
-
Subscription:
-
Upgrade: You must pay the price difference between the new and old specifications for the remaining subscription period.
-
Downgrade: The system automatically calculates the remaining value and issues a refund. For more information, see Refund rules for downgrades.
-
Related APIs
|
API |
Description |
|
Modifies the specifications of a single node in a PolarDB cluster. |
|
|
Modifies the specifications of multiple nodes in a PolarDB cluster. |