PolarDB lets you manage individual nodes. This topic describes common node operations.
Rename a node
If your cluster uses multiple nodes for different business workloads, you can assign them custom names for easier identification in the console.
Usage notes
A node and its name are linked. During a failover, node names are not swapped; only the role of each node changes. For more information, see Automatic and manual failover.
-
Before the failover, the node named RW has the primary node role, and the node named RO1 has the read-only node role.
-
After the failover:
The node names remain the same, but their role is swapped. The node originally named RW now has the read-only node role, and the node originally named RO1 has the primary node role.
Procedure
Go to the page and find your target cluster. On the cluster's Basic Information page, go to the Database Nodes section and find the node that you want to rename. Click Modify next to the Node Name.
Restart a node
You can manually restart a node in the console to resolve database connectivity or performance issues. In addition, PolarDB provides a serverless feature that allows nodes in your cluster to scale elastically within seconds.
Usage notes
-
When a read-only node is restarted, new read/write splitting connections are routed to it. Existing connections are not routed to the restarted node until they reconnect. To force a reconnection, restart your application.
-
Restarting a node can cause a brief service interruption of up to one minute. We recommend performing this operation during off-peak hours and ensuring that your application has a reconnection mechanism.
-
The time required to restart a node depends on the data volume and can take several hours. Proceed with caution.
Procedure
Go to the page and find your target cluster. On the cluster's Basic Information page, go to the Database Nodes section and find the node that you want to restart. In its information box, click Restart.
API reference
|
API |
Description |
|
Restarts a node of a PolarDB cluster. |