DEV Community

Chen Debra
Chen Debra

Posted on

Apache DolphinScheduler 3.3.2 Released! Major Updates in Performance and Stability

#apachedolphinscheduler #release #opensource #bigdata

3.3.2

We’re excited to announce that Apache DolphinScheduler 3.3.2 has been officially released!

This update brings a range of performance improvements, stability enhancements, documentation updates, and critical bug fixes, making workflow orchestration even smoother and more reliable for users.

Key Improvements

1. Enhanced Stability and Database Performance

  • Added index on workflow_definition_code in table t_ds_schedules to greatly reduce database query time when accessing schedules. (#17513 by @unigof)
  • Fixed potential NPE when handling Zookeeper connection events, improving overall service resilience. (#17526 by @Mrhs121)

2. Master Node Optimization

  • The default value of batchTriggerAcquisitionMaxCount is now aligned with threadCount, ensuring balanced load and smoother scheduling. (#17483 by @ruanwenjun)
  • Added support for separate data sources for Quartz, providing more flexible configurations for enterprise deployments. (#17468 by @ruanwenjun)

3. Improved Plugin and Storage Management

  • The local storage implementation has been decoupled from the HDFS plugin, enabling more modular and lightweight deployment options. (#17547 by @ruanwenjun)
  • Fixed multiple issues related to HDFS storage type startup and mounting paths in Kubernetes environments. (#17496 by @SbloodyS, #17517 by @cn-hew)

4. Documentation and Configuration Enhancements

  • Optimized deployment documentation and fixed incorrect document locations of DolphinDB, etc. (#17491 by @SbloodyS, #17444 by @SbloodyS)
  • Removed outdated task definition docs to keep content concise and relevant. (#17448 by @SbloodyS)
  • Improved CI configurations and removed unused dependencies (such as zt-zip) to streamline the build process. (#17525 by @ruanwenjun)

Major Bug Fixes

  • SQL Task parameter passing issue resolved — users can now pass parameters correctly in SQL tasks. (#17456 by @Zzih96)
  • Fixed workflow deletion bug where a workflow containing failover instances could still be removed. (#17478 by @ruanwenjun)
  • Fixed TASK_ONLY strategy issue in workflow execution strategy. (#17461 by @ruanwenjun)
  • Aliyun SS Task final state bug fixed, ensuring proper task lifecycle completion. (#17475 by @EricGao888)
  • ThreadLocal cleanup improved to prevent leaks when exceptions occur in the login interceptor. (#17474 by @njnu-seafish)
  • Fixed Hive & Spark data source principal field visibility and correctness under Kerberos environments. (#17493 by @njnu-seafish)
  • Duplicate task name validation added when saving or updating workflows. (#17576 by @njnu-seafish)
  • Fixed variable display issue when setting startup parameters for workflow instances. (#17583 by @Mrhs121)
  • Fixed task dispatch blocking caused by high-priority delay events. (#17556 by @ruanwenjun)
  • Fixed sub-workflow scheduling issue in API layer. (#17549 by @shangeyao)

These fixes collectively enhance workflow reliability, task execution consistency, and system robustness in distributed environments.

Chore & CI Improvements

Continuous integration (CI) and repository optimizations remain a focus:

  • Fixed several flaky CI tests and deadlink validation issues.
  • Improved POM configuration and changed module dependency scopes to “provided” where appropriate.
  • Updated project version to 3.3.2 and cleaned unused libraries for a more efficient build process.

A Big Thank You to Our Contributors!

This release wouldn’t be possible without the dedication and hard work of our community contributors:

@Gallardot, @Mrhs121, @SbloodyS, @ruanwenjun, @njnu-seafish, @cn-hew, @EricGao888, @shangeyao, @unigof, @LourierL, @Zzih96

Your contributions — from core code improvements to documentation fixes and CI maintenance — continue to make Apache DolphinScheduler more stable, powerful, and user-friendly.

What’s Next

The community is actively working toward 3.3.3 and 4.0 milestone features, focusing on:

  • Workflow performance optimization for ultra-large-scale scenarios
  • Task plugin refactoring to improve extensibility
  • Enhanced observability and scheduling intelligence

Stay tuned for more exciting updates — and as always, we welcome your participation!
You can:

Top comments (0)