Multiple compute nodes are deployed in the primary zone of a PolarDB cluster of Cluster. These compute nodes consist of one primary node and one or more read-only nodes.

Limits

Only PolarDB clusters of Cluster and Archive Database Cluster Edition support multi-node deployment. Single Node and Archive Database (High Compression Ratio) do not support this feature. For more information, see Editions.

Multi-node architecture

1

Each PolarDB cluster contains one primary node and one or more read-only nodes. A cluster can contain at most 15 read-only nodes and must contain at least one read-only node. All of the nodes in the same cluster use the same specifications.

The multi-node architecture ensures the high availability of PolarDB clusters. When the primary node in a cluster fails, the cluster can automatically fail over to a read-only node. Then, the read-only node serves as the new primary node.

In addition, PolarProxy of a PolarDB cluster can be used to perform read/write splitting. For more information, see PolarProxy.

Add or remove read-only nodes

A PolarDB cluster of Cluster Edition can contain at most 15 read-only nodes. Each cluster must contain at least one read-only node to ensure high availability. All nodes in a cluster use the same specifications.

You can manually add or remove read-only nodes to adjust the cluster performance based on your requirements. For more information, see Add or remove read-only nodes.

Note
  • The system requires about 5 minutes to add a read-only node. The time consumption depends on multiple factors, such as the number of newly added nodes, the numbers of databases and tables, and the database loads. When PolarDB adds nodes to a cluster, the databases in the cluster are not affected.
  • When PolarDB removes a read-only node from a cluster, connections to the node are closed. The connections to other nodes in the cluster are not affected. We recommend that you remove nodes during off-peak hours and make sure that your applications are configured with the automatic reconnection mechanism.
  • We recommend that you connect your application to the cluster endpoint because PolarDB can automatically re-create connections after you add or remove read-only nodes. This way, you no longer need to modify the application configurations. After you add a read-only node, PolarDB automatically discovers the node and then balances traffic loads to the node. After you remove a read-only node, PolarDB automatically filters out the node.