DBLE 3.21.02.0 Release Notes Overview

Version 3.21.02.0 – DBLE Release Notes

This article provides a detailed interpretation of the DBLE 3.21.02.0 release notes, organized into four main sections: project introduction, new version feature analysis, complete release notes, and LTS version updates.

1. DBLE Project Introduction

DBLE is an enterprise‑grade open‑source distributed middleware, nicknamed “MyCat Plus”. It is praised for its simplicity, stability, continuous maintenance, vibrant community, and broad user base.

Key information:

Project start date: 2017‑10‑24

Focus: Open‑source

Capabilities: Data horizontal sharding, read/write splitting, distributed transaction support, multi‑shard algorithms, global ID, IP/SQL whitelist‑blacklist

Strengths: MySQL syntax compatibility, complex query optimization, low migration cost, mature and stable technology stack

Official resources: GitHub: https://github.com/actiontech/dble Documentation: https://actiontech.github.io/dble-docs-cn Releases: https://github.com/actiontech/dble/releases

2. New Version Feature Highlights

The 3.21.02.0 release is the first DBLE release of the year, adding and refactoring ten features, fixing more than ten defects, and making two backward‑compatibility adjustments.

1) General Log

Enabling the general log records every SQL statement received by DBLE to a file, with an estimated performance overhead of 3%–5%. It supports configurable log directory and filename, file size rotation, hot‑switching, and management commands to view current settings.

2) Load Data Batch Mode

When using load data , network timeouts previously caused full rollbacks. The new batch mode splits the import file into multiple chunks (as defined in bootstrap.cnf ), allowing successful chunks to remain while stopping on the failing chunk. Errors in individual files are reported without aborting the entire operation.

3) Fine‑Grained SQL Statistics

DBLE can now collect statistics on transactions and CRUD operations per table, user, and front‑end/back‑end dimensions, including execution count, latency, and row counts. The feature can be toggled via a hot‑switch and threshold settings.

3. Complete Release Notes

Features

[#2409] Support for general log.

[#2428] Additional SQL statistics.

[#2387] Server‑side cursor support.

[#2503] New rwSplitMode.

[#2386] Internal JSON configuration mode.

[#1830] Display and resume paused data nodes in a cluster.

[#2351] Show “dble process list” on server ports.

[#2595] Load data batch processing.

[#2594] Upgrade Druid to version 1.2.3.

Major Bug Fixes

[#1843] Correct error handling when disabling readHost then writeHost on the same host.

[#2497] Fix ArrayIndexOutOfBoundsException in slow query log.

[#2377] Resolve conflict between show @@ processlist and RwSplitService.

[#2199] Correct execution of AND/OR statements.

[#2152] Route columns without explicit table names correctly.

[#2342] Ensure shardingNode uniqueness in table configuration.

[#2366] Clean up stray “/” in error messages during config reload.

[#2105][#2389] Navicat connection issues resolved.

[#2034] Fix “Union field merge error, field not found: W.SCANNUM”.

[#2324] Show file causing dry‑run errors.

Backward‑Compatibility Breaks

[#2386] Internal JSON configuration mode changes.

[#2598] homePath must be explicitly set in bootstrap.cnf when upgrading.

4. LTS Version Updates

Minor LTS releases address bug fixes:

3.20.07.3 – Fixed connection pool parameter errors, alert issues, SELECT…FOR UPDATE routing, and DDL execution without a dbgroup.

3.20.10.2 – Same fixes as above plus Zookeeper upgrade to 3.4.12.

Community acknowledgments thank contributors @evangoe, @Neil‑Zhou, @314926481, @ClippedWings‑renzy, and the broader GitHub and QQ community for issues, stars, and feedback.

For more details, refer to the official DBLE documentation, release pages, and the linked tutorial videos.