Skip to content

[Failing Test]: Flink 1.20 portable DataSet ValidatesRunner tests hang #40645

Closed

Description

@aIbrahiim

What happened?

Flink 1.20 portable DataSet ValidatesRunner jobs were hanging: pipelines stayed RUNNING until waitUntilFinish hit the 20 min timeout. Already happening on master.

Examples that hung:

  • CombineTest$WindowingTests.testWindowedCombineGloballyAsSingletonView
  • FlattenTest.testFlattenPCollections
  • MapViewTest.testMapSideInput
  • ParDoTest$MultipleInputsAndOutputTests.testSameSideInputReadTwice
  • ParDoTest$TimerTests.testEventTimeTimerBounded
  • SplittableDoFnTest.testWindowedSideInputBounded
  • ViewTest.testEmptyListSideInput

https://github.com/apache/beam/actions/runs/38961839107/job/117178555850

Root cause was portable stage contexts sticking around across jobs in the same JVM. Fixed in #40632 (per job cleanup in FlinkExecutableStageContextFactory + portable VR parallelThreads=1).

Issue Failure

Failure: Test is continually failing

Issue Priority

Priority: 1 (unhealthy code / failing or flaky postcommit so we cannot be sure the product is healthy)

Issue Components

  • Component: Python SDK
  • Component: Java SDK
  • Component: Go SDK
  • Component: Typescript SDK
  • Component: IO connector
  • Component: Beam YAML
  • Component: Beam examples
  • Component: Beam playground
  • Component: Beam katas
  • Component: Website
  • Component: Infrastructure
  • Component: Spark Runner
  • Component: Flink Runner
  • Component: Prism Runner
  • Component: Twister2 Runner
  • Component: Hazelcast Jet Runner
  • Component: Google Cloud Dataflow Runner

Activity

  1. aIbrahiim commented on Oct 17, 2026

    @aIbrahiim
    ContributorAuthor
  2. schizophrenicmaniac commented on Oct 19, 2026

    @schizophrenicmaniac
    Contributor

    .take-issue

    Hi, I'd like to work on this. My first suspect is #39663, which landed just before this was reported and changed how the Java harness shuts down its gRPC channels. I'd reproduce with validatesPortableRunnerBatchDataSet on 1.20, fix the hang, and re-enable the tests skipped in #40632.

  3. aIbrahiim commented on Oct 19, 2026

    @aIbrahiim
    ContributorAuthor

    4. .take-issue
    Hi, I'd like to work on this. My first suspect is #39663, which landed just before this was reported and changed how the Java harness shuts down its gRPC channels. I'd reproduce with validatesPortableRunnerBatchDataSet on 1.20, fix the hang, and re-enable the tests skipped in #40632.

    Hi @schizophrenicmaniac, thank you but those skips aren't in #40632 anymore and the hang was stage contexts sticking around across jobs in the same jvm, fixed that in the flink factory and PostCommit Flink Batch is green with the DataSet suite running again so i don't think #39663 is the cause here

  4. added this to the 2.78.0 Release milestone on Oct 21, 2026
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Metadata

Metadata

Type

No type

Projects

No projects

    Relationships

    None yet

    Development

    No branches or pull requests

    Issue actions