luoyuan3471 opened a new pull request, #8915:
URL: https://github.com/apache/hadoop/pull/8915
## Problem
- `RegularContainerAllocator.assignContainer` logs a WARN when a request
exceeds the candidate node's total resource capacity.
- In heterogeneous clusters, a valid request may fit larger nodes but not
smaller ones. Repeated attempts on smaller nodes can therefore generate
WARN noise during normal scheduling.
## Solution
- Change this message from WARN to parameterized DEBUG logging.
- Preserve `LOCALITY_SKIPPED` and existing scheduling activity diagnostics.
- Leave resource checks, allocation decisions, and other warning messages
unchanged.
- Add a regression test covering INFO/DEBUG logging, skipping an undersized
node without reserving, and allocation on a larger node.
## Validation
- Local JDK 17 and Maven Wrapper 3.9.15: ResourceManager compilation passed
using locally installed dependencies. No native profile was enabled.
- `TestContainerAllocation`: 18 tests, 0 failures, 0 errors, 0 skipped.
- With the original production code, the new regression test failed at the
INFO-level log assertion: 1 failure, 0 errors, 0 skipped.
- Fresh Surefire XML and `git diff --check` verified. Community CI has not
run.
No benchmark result is claimed.
- Contains content generated by Codex.
## Related Jira
[YARN-12058](https://issues.apache.org/jira/browse/YARN-12058)
--
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]
