-
Notifications
You must be signed in to change notification settings - Fork 4.5k
Description
What happened?
Run the TestXlang_CoGroupBy test from the Go validates runner suite.
Data gets corrupted, and if the Go SDK does loopback execution the passert returns the following:
actual PCollection does not match expected values
=========
0 correct entries (present in both)
=========
2 unexpected entries (present in actual, missing in expected)
+++
1964988:[]
+++
1:[]
=========
2 missing entries (missing in actual, present in expected)
---
0:[1 2 4]
---
1:[3 5 6]
As this is only happening in prism (other runners seem to cope), either prism's unfused execution is causing corruption on the Java side, OR it's prism receiving it is causing the corruption. Either way, it needs to be fixed so the pipeline passes.
The test in question is here:
| func TestXLang_CoGroupBy(t *testing.T) { |
It needs to be unfilered from the prism run here though:
beam/sdks/go/test/integration/integration.go
Line 143 in 228157f
| "TestXLang_CoGroupBy", |
The following action workflow runs the suite:
https://github.com/apache/beam/actions/workflows/beam_PreCommit_GoPrism.yml
Issue Priority
Priority: 2 (default / most bugs should be filed as P2)
Issue Components
- Component: Python SDK
- Component: Java SDK
- Component: Go SDK
- Component: Typescript SDK
- Component: IO connector
- Component: Beam examples
- Component: Beam playground
- Component: Beam katas
- Component: Website
- Component: Spark Runner
- Component: Flink Runner
- Component: Samza Runner
- Component: Twister2 Runner
- Component: Hazelcast Jet Runner
- Component: Google Cloud Dataflow Runner