markap14 commented on PR #11164: URL: https://github.com/apache/nifi/pull/11164#issuecomment-4443838233
[claude-opus-4.7] CI rerun on `macos-15 Java 25` failed again, but on a **different** test than before. Classifying as a separate likely flake and rerunning the shard once more. **Failed check:** `macos-15 Java 25` — https://github.com/apache/nifi/actions/runs/25806807059/job/75830543159 **Failing test (this run):** `org.apache.nifi.tests.system.parameters.ClusteredParameterContextIT > ParameterContextIT.testProcessorRestartedAfterLongDependentServiceValidationPeriod` ``` [ERROR] org.apache.nifi.tests.system.parameters.ClusteredParameterContextIT.testProcessorRestartedAfterLongDependentServiceValidationPeriod -- Time elapsed: 15.54 s <<< ERROR! org.apache.nifi.toolkit.client.NiFiClientException: Error deleting Controller Service: Node localhost:5672 is unable to fulfill this request due to: StandardControllerServiceNode[service=StandardSleepService[...], name=StandardSleepService, active=false] cannot be deleted because it is not disabled Caused by: jakarta.ws.rs.ClientErrorException: HTTP 409 Conflict ``` **Why this is most likely unrelated to the PR:** - The `macos-15 Java 25` shard is independently flaky on this PR — the previous rerun was for a completely different test (`ClusterStateKeyDropIT.testCanDropSpecificStateKey`). Same shard failing on a different test each time is the signature of an environmentally flaky runner, not a deterministic regression from this PR. - All other shards pass: `ubuntu-24.04 Java 21`, `ubuntu-24.04 Java 25`, `macos-15 Java 21`, plus all the language/locale shards. - The test contains a pre-existing race independent of this PR: `NiFiClientUtil.disableControllerService(...)` (`nifi-system-test-suite/src/test/java/org/apache/nifi/tests/system/NiFiClientUtil.java:1014-1021`) only fires the disable request and returns immediately. It does **not** wait for the service to reach the `DISABLED` state. The very next line in the test (`ParameterContextIT.java:614-615`) calls `deleteControllerService`. The `StandardSleepService` under test has `Validate Sleep Time = 6 secs` set via the parameter context update, so if a validation cycle is in flight when disable is issued, the service stays in `DISABLING` long enough for the immediate delete to fail with HTTP 409 — exactly the observed error. - This PR does not change controller-service lifecycle, validation scheduling, or the `disableControllerService` helper. The scheduling changes only affect processor / reporting-task `onTrigger` invocations. **Action taken:** Reran only the failed `macos-15 Java 25` shard a second time via `gh run rerun 25806807059 --failed --repo apache/nifi`. If the same test fails again on the next run, I will stop reruns and investigate whether virtual-thread scheduling can prolong validation on macOS Java 25 specifically. -- 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]
