HyukjinKwon commented on PR #57017: URL: https://github.com/apache/spark/pull/57017#issuecomment-4888037942
Independent corroboration (another agent in the same multi-agent session): I reached the identical root cause and fix, and ran a harder repeat-validation to account for the intermittent (~25-50%) nature of the original flake. **6 consecutive full `YarnClusterSuite` runs on JDK 17, all green** (run [28757813841](https://github.com/HyukjinKwon/spark/actions/runs/28757813841)): | rep | result | |-----|--------| | 1 | ✅ success | | 2 | ✅ success | | 3 | ✅ success | | 4 | ✅ success | | 5 | ✅ success | | 6 | ✅ success | The six formerly-timing-out cluster-mode tests (`displays driver log links`, `timeout to get SparkContext`, `executor env overwrite AM env` client/cluster, `SPARK-34472 ivySettings` ×2) pass in every rep. I verified from the failing run's artifact that the mini NM was advertising only `<memory:4096>` (the `yarn.minicluster.*` default) despite SPARK-57710 setting `yarn.nodemanager.resource.memory-mb=8192` — exactly as this PR describes. LGTM. -- 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]
