joseluisll commented on PR #8681: URL: https://github.com/apache/hadoop/pull/8681#issuecomment-5485433056
The tests in the `-1 unit` fail on clean trunk as well. I ran `TestRouterWebServicesREST` three times on unpatched trunk, at the same revision this precommit used (`abbc78d4416`): | | failures / 42 | |---|:--:| | trunk, run 1 | 2 | | trunk, run 2 | 1 | | trunk, run 3 | 3 | | this patch | 1 | Same code, different results each run, always among `testAppAttemptXML`, `testAppXML` and `testGetAppsMultiThread`. Trunk fails more often than the patch. The cause is a race in the test: `getAppAttempt` reads `getAttempts().get(0)` right after `submitApplication()` without waiting, so if the RM has not created the attempt yet the list is empty. That is the `IndexOutOfBounds: Index 0 out of bounds for length 0` in the precommit log. `TestFederationWebApp` extends that class and fails the same way. `TestYarnFederationWithFairScheduler`, the third name in the report, passes 38/38. The suite this JIRA fixes, `TestYarnFederationWithCapacityScheduler`, passes 38/38. Its child JVMs run the shutdown hooks to completion, stop the token remover and the SubClusterCleaner, and close every ZooKeeper session they open. -- 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]
