pan3793 opened a new pull request, #8549:
URL: https://github.com/apache/hadoop/pull/8549
<!--
Thanks for sending a pull request!
1. If this is your first time, please read our contributor guidelines:
https://cwiki.apache.org/confluence/display/HADOOP/How+To+Contribute
2. Make sure your PR title starts with JIRA issue id, e.g.,
'HADOOP-17799. Your PR title ...'.
-->
### Description of PR
Fix flaky tests like
https://github.com/kokonguyen191/hadoop/actions/runs/27467848524/job/81194100665
```
Error: Errors:
Error: org.apache.hadoop.metrics2.sink.TestPrometheusMetricsSink.testPublish
Error: Run 1: TestPrometheusMetricsSink.testPublish:60 » Metrics Metrics
source TestMetrics already exists!
Error: Run 2: TestPrometheusMetricsSink.testPublish:60 » Metrics Metrics
source TestMetrics already exists!
Error: Run 3: TestPrometheusMetricsSink.testPublish:60 » Metrics Metrics
source TestMetrics already exists!
[INFO]
Error:
org.apache.hadoop.metrics2.sink.TestPrometheusMetricsSink.testPublishFlush
Error: Run 1: TestPrometheusMetricsSink.testPublishFlush:159 The first
metric should not exist after flushing ==> expected: <false> but was: <true>
Error: Run 2: TestPrometheusMetricsSink.testPublishFlush:137 » Metrics
Metrics source TestMetrics already exists!
Error: Run 3: TestPrometheusMetricsSink.testPublishFlush:137 » Metrics
Metrics source TestMetrics already exists!
[INFO]
Error:
org.apache.hadoop.metrics2.sink.TestPrometheusMetricsSink.testPublishMultiple
Error: Run 1: TestPrometheusMetricsSink.testPublishMultiple:112 The
expected first metric line is missing from prometheus metrics output ==>
expected: <true> but was: <false>
Error: Run 2: TestPrometheusMetricsSink.testPublishMultiple:95 » Metrics
Metrics source TestMetrics1 already exists!
Error: Run 3: TestPrometheusMetricsSink.testPublishMultiple:95 » Metrics
Metrics source TestMetrics1 already exists!
[INFO]
[INFO]
Error: Tests run: 5385, Failures: 0, Errors: 3, Skipped: 201
```
The root cause is the refCount-based global singleton:
- `DefaultMetricsSystem.instance()` returns a JVM-global singleton shared by
all tests in hadoop-common.
- `init()` short-circuits (returns early without incrementing refCount) if
monitoring is already `true`.
- `shutdown()` only clears `allSources`/`allSinks` when refCount hits 0.
So if any test (this class's own methods, or any other metrics test in the
module) throws before its inline `stop()`/`shutdown()`, the singleton stays
`monitoring=true` with leaked sources. Subsequent tests' `init()` becomes a
no-op, sources never get cleared, and you get cascading "Metrics source
TestMetrics already exists!" plus stale-data assertion failures — exactly what
the CI shows (including on surefire reruns).
### How was this patch tested?
Pass GHA.
### For code changes:
- [x] Does the title or this PR starts with the corresponding JIRA issue id
(HADOOP-19920)?
- [ ] Object storage: have the integration tests been executed and the
endpoint declared according to the connector-specific documentation?
- [ ] If adding new dependencies to the code, are these dependencies
licensed in a way that is compatible for inclusion under [ASF
2.0](http://www.apache.org/legal/resolved.html#category-a)?
- [ ] If applicable, have you updated the `LICENSE`, `LICENSE-binary`,
`NOTICE-binary` files?
### AI Tooling
Contains content generated by Claude Opus 4.8
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: [email protected]
For queries about this service, please contact Infrastructure at:
[email protected]
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
