Caideyipi opened a new pull request, #19022:
URL: https://github.com/apache/iotdb/pull/19022

   ## Description
   
   
`PipeProcessorSubtaskWorkerTest.testOneHundredWaitingPipesDoNotWaitForWorkerPolling`
 requires 100 waiting pipes to finish within five seconds. A correctly notified 
worker delayed by CI scheduling or a JVM pause can exceed that deadline and 
fail the admission-notification assertion.
   
   Disable idle polling with `Object.wait(0)` and wait for workers to reach 
`WAITING`. Allow 30 seconds for completion within a 60-second JUnit timeout. 
Missing notifications still leave workers blocked, while the completion 
deadline allows more scheduling headroom.
   
   ## Validation
   
   - `mvn spotless:apply -P with-integration-tests -pl iotdb-core/datanode` 
passed.
   - A clean reactor build of DataNode and its IoTDB dependency closure, with 
Develocity caching disabled, passed all 23 tests in 
`PipeProcessorSubtaskWorkerTest`, `PipeMemoryManagerTest`, 
`PipeProcessorSubtaskExecutionGuardTest`, and 
`PipeProcessorSubtaskExecutorTest`.
   - Repeated the complete worker test class 20 times on one CPU: 120 tests 
passed.
   - Injected a six-second delay into one worker in a temporary test variant: 
the original test failed with the reported admission-notification assertion; 
the updated test passed.
   - Suppressed worker notifications in a temporary negative control: the 
updated test failed with the admission-notification assertion as expected.
   
   The full unit-test suite was not run. Temporary experiment code is excluded 
from this PR.
   
   - [x] Self-reviewed the complete diff.
   - [x] Updated the existing concurrency regression test and explained its 
waiting strategy.
   


-- 
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]

Reply via email to