luoyuan3471 opened a new pull request, #8854: URL: https://github.com/apache/hadoop/pull/8854
### Description of PR When keepContainersAcrossApplicationAttempts is enabled, a container from an earlier application attempt may remain in the scheduler after the old attempt has been removed from RMApp. ClientRMService#getContainerReport currently rejects this request before consulting the scheduler. This patch skips the application-attempt existence check only for applications that retain containers across attempts. Application ACL validation and the scheduler-side container existence check remain unchanged. JIRA: https://issues.apache.org/jira/browse/YARN-12039 ### How was this patch tested? Added regression coverage for: - returning a report for a retained container from a removed attempt; - preserving ApplicationAttemptNotFoundException when container transfer is disabled; - preserving ContainerNotFoundException when the scheduler no longer contains the container. The three focused tests pass, and the complete TestClientRMService class passes (43 tests). ### For code changes: - [x] Does the title of this PR start with the corresponding JIRA issue id? - [ ] Object storage: Not applicable. - [ ] No new dependencies were added. - [ ] No LICENSE or NOTICE changes are required. -- 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]
