All Products
Search
Document Center

PolarDB:Database proxy FAQ

Last Updated:Oct 08, 2026

This topic provides answers to frequently asked questions about the PolarDB database proxy.

Why can't I read data immediately after inserting it?

In a read/write splitting architecture, replication between the primary node and the read-only nodes causes latency. However, PolarDB supports session consistency, which guarantees that reads within the same session reflect your previous writes. For more information, see Session consistency.

Are zero-latency reads supported?

PolarDB clusters have millisecond-level replication latency between the primary node and the read-only nodes under normal load. Read/write splitting endpoints do not support zero-latency reads immediately after a write. If you require zero-latency reads, send both read and write requests to the primary endpoint, which dynamically points to the PolarDB primary node. To view the primary endpoint, see View endpoints and ports.

Why is the load on the primary node high?

High CPU load on the primary node can have multiple causes. Check the following common items:

  • If your application accesses the primary endpoint directly, check whether any services are configured to use the primary endpoint.

  • Your workload sends far more write requests than read requests, which places heavy load on the primary node.

  • If the proxy endpoint has Primary Node Accepts Read Requests set to Yes, change Primary Node Accepts Read Requests to No to reduce the number of read requests that are routed to the primary node.

  • Check whether Transaction Splitting is enabled for the proxy endpoint. Some ORMs wrap all requests in transactions by default, which routes all requests to the primary node. Enable Transaction Splitting to reduce the number of read-before-write requests in transactions that are routed to the primary node.

If you run Sysbench stress tests, add --oltp-skip-trx=on in Sysbench 0.5, or add --skip-trx=on in Sysbench 1.0 to disable transactions before you run the test. For more information about load balancing policies and transaction splitting, see Load balancing.

Why does one node receive more requests than other nodes?

Requests are distributed based on node load. A node with a lower load receives more requests.

Are newly added read-only nodes automatically added to read/write splitting?

Yes.

Does setting the consistency level to global consistency (strong) cause a performance difference compared with eventual consistency (weak) and session consistency (medium)?

Global consistency (strong) causes some performance overhead compared with eventual consistency (weak) and session consistency (medium). If your workload requires high concurrency, change the consistency level during off-peak hours.

What is the difference between Standard Enterprise Edition and Dedicated Enterprise Edition in PolarProxy Enterprise Edition?

  • Standard Enterprise Edition: corresponds to the General-purpose cluster sub-edition. It shares physical CPU resources and provides intelligent, second-level elastic scaling based on workload.

  • Dedicated Enterprise Edition: corresponds to the Dedicated cluster sub-edition. It uses dedicated physical CPU resources for better performance stability.

Does removing a selected read-only node from a cluster endpoint affect my workloads?

  • If you remove a read-only node from a cluster endpoint in read/write splitting mode, the workload connections on that endpoint are disconnected only when the removed node has requests, transactions, or temporary tables in progress. Otherwise, your workloads are not affected.

  • If you remove a read-only node from a cluster endpoint in read-only mode, the workload connections on that endpoint are disconnected.

For more information about the impact of removing a read-only node, see Read-only node removal.

What read/write modes does PolarDB support?

Currently, PolarDB cluster endpoints and custom endpoints support two read/write modes: Read-only and Read/Write (automatic read/write splitting).

  • Read-only mode: All requests are routed only to read-only (RO) nodes. The read/write (RW) node cannot be attached to a read-only endpoint. If a workload attempts to write to a read-only endpoint, an error is returned. This mode supports load balancing based on the number of connections and the number of active requests.

  • Read/write splitting mode: Read/write splitting is automatically enabled. By default, the read/write (RW) node is attached. This mode supports load balancing based only on the number of active requests.

Changing the read/write mode takes effect only for new connections. Existing connections keep the previous mode. For more information about the configuration, see Configure the database proxy.

Why does a binary log dump return the error "Connections-based load balancing policy does not support BINLOG_DUMP in non-block mode"?

Currently, read-only endpoints that use the connections-based Load balancing load balancing policy do not support pulling binary log files in non-block mode. In the console, change the load balancing policy of the read-only endpoint to load balancing based on the number of active requests.

How are database proxy versions 1.x.x and 2.x.x determined?

Currently, the PolarDB for MySQL database proxy has two major versions: 1.x.x and 2.x.x. The differences between the two versions are as follows:

  • 1.x.x

    The database proxy version for clusters created before February 1, 2021. This version no longer receives new features and receives only bug fixes.

  • 2.x.x

    The database proxy version for clusters created on or after February 1, 2021. This is the current mainstream version, and all new features are developed on it. New features include persistent connections and data masking.

Can I upgrade the database proxy from 1.x.x to 2.x.x by performing a minor version upgrade in the console?

No. Database proxy version 1.x.x can be upgraded only to the latest version within the 1.x.x series. It cannot be upgraded directly to 2.x.x. If you have this requirement, see Can the database proxy be upgraded from 1.x.x to 2.x.x?.

Can the database proxy be upgraded from 1.x.x to 2.x.x?

Yes, but your cluster must meet the following requirements:

  • If your cluster runs PolarDB for MySQL 8.0.2, there is no minor version requirement.

  • If your cluster runs PolarDB for MySQL 8.0.1, the revision version must be 8.0.1.1.10 or later.

  • If your cluster runs PolarDB for MySQL 5.7, the revision version must be 5.7.1.0.6 or later.

  • If your cluster runs PolarDB for MySQL 5.6, the revision version must be 5.6.1.0.24 or later.

If your cluster meets the preceding requirements, contact us for assistance.

Is the database proxy automatically upgraded to 2.x.x during a major engine version upgrade in the console?

Yes. During a major engine version upgrade, the database proxy is automatically upgraded to version 2.x.x. You cannot specify version 1.x.x.

Why do connection errors occur after I set the weight of a read-only node to 0 in a cluster endpoint and then remove the node?

  • Scenario: You want to remove a read-only node from a cluster endpoint. To avoid affecting the connections on that node, you first set the read/write splitting weight of the node to 0. You expect that subsequent read requests are no longer routed to the node. However, after you remove the node from the cluster endpoint, your application reports connection errors.

  • Cause: A cluster endpoint supports two types of Load Balancing Policy, namely Active Request-based Load Balancing and Connections-based Load Balancing:

    • Active Request-based Load Balancing: After you reduce the weight of a read-only node to 0, subsequent requests are no longer routed to that node.

    • Connections-based Load Balancing: Load balancing is performed only when a connection is established. If an established connection is routed to a read-only node, all subsequent requests over that connection continue to be sent to that node.

  • Solution:

    1. Change the Load Balancing Policy of the cluster endpoint to Active Request-based Load Balancing.

    2. Set the read/write splitting weight of the target read-only node to 0.

    3. Remove the read-only node.