FrankChen021 opened a new pull request, #20096: URL: https://github.com/apache/druid/pull/20096
## Summary - Reduce the faulty worker's Overlord task-status delay in `MSQWorkerFaultToleranceTest` from `PT1H` to `PT1S`. - The test still uses the faulty Indexer, worker capacity of one, explicit worker cancellation, failed-task metric, fault-tolerant relaunch, segment availability, and final SQL result assertion. ## Evidence The `M*,P*,O*` CI shard in run [#20087](https://github.com/apache/druid/actions/runs/32339940988?pr=20087) reported this test as the shard's slowest test: - `MSQWorkerFaultToleranceTest#test_cancelledWorker_isRetried_ifFaultToleranceIsEnabled`: 301.197s - Next slowest test: `MergingRowIteratorTest#testAllPossible5ElementSequences`: 69.853s - The 301.197s test accounts for 25.5% of the shard's 1,179.924 recorded testcase-seconds. `FaultyOverlordClient` sleeps for the configured delay on every `taskStatus`/`taskStatuses` call. The test only needs to hold the faulty worker long enough to observe `ingest/count` and cancel it; the one-hour delay was a blocking-wait bottleneck. ## Local measurements Command used for each isolated run: ```text mvn -ntp test -pl embedded-tests -am \ -Dtest=org.apache.druid.testing.embedded.msq.MSQWorkerFaultToleranceTest#test_cancelledWorker_isRetried_ifFaultToleranceIsEnabled \ -Dsurefire.failIfNoSpecifiedTests=false -Pskip-static-checks -Dweb.console.skip=true -T1C ``` | Variant | Passing runs | Surefire | Maven wall | | --- | ---: | ---: | ---: | | Unmodified `PT1H` | 1 | 312.00s | 353.09s | | `PT1S` | 1 | 13.07s | 38.71s | | `PT1S` | 2 | 12.62s | 37.96s | Optimized average: 12.845s Surefire and 38.335s Maven wall. - Surefire saving: 299.155s (95.88%). - Maven wall saving: 314.755s (89.14%). - Optimized Surefire range: 12.62–13.07s (0.45s spread). The complete affected test class passed in 12.57s. ## Validation - `mvn -ntp verify -pl embedded-tests -am -DskipTests -Dweb.console.skip=true -T1C` - `mvn -ntp com.github.spotbugs:spotbugs-maven-plugin:4.10.2.0:check -pl embedded-tests -DskipTests -Dweb.console.skip=true` - `git diff --check` All passed. This is part of #13948. -- 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]
