[FLINK-13009] YARNSessionCapacitySchedulerITCase#testVCoresAreSetCorrectlyAndJobManagerHostnameAreShownInWebInterfaceAndDynamicPropertiesAndYarnApplicationNameAndTaskManagerSlots throws NPE on Travis

• Type: Improvement
• Status: Closed
• Priority: Major
• Resolution: Duplicate
• Affects Version/s: 1.8.0
• Component/s: Deployment / YARN, Tests
• Labels: test-stability
• Assignee: Unassigned
• Reporter: Congxian Qiu
• Created: 27/Jun/19 03:36
• Updated: 27/Apr/21 18:08
• Resolved: 02/Jun/20 10:19

Description

The test YARNSessionCapacitySchedulerITCase#testVCoresAreSetCorrectlyAndJobManagerHostnameAreShownInWebInterfaceAndDynamicPropertiesAndYarnApplicationNameAndTaskManagerSlots throws NPE on Travis.

NPE throws from RMAppAttemptMetrics.java#128, and the following is the code from hadoop-2.8.3[1]

// Only add in the running containers if this is the active attempt.
128	RMAppAttempt currentAttempt = rmContext.getRMApps()
129	.get(attemptId.getApplicationId()).getCurrentAppAttempt();

 

log https://api.travis-ci.org/v3/job/550689578/log.txt

[1] https://github.com/apache/hadoop/blob/b3fe56402d908019d99af1f1f4fc65cb1d1436a2/hadoop-yarn-project/hadoop-yarn/hadoop-yarn-server/hadoop-yarn-server-resourcemanager/src/main/java/org/apache/hadoop/yarn/server/resourcemanager/rmapp/attempt/RMAppAttemptMetrics.java#L128

Issue Links

• is duplicated by: FLINK-15534 YARNSessionCapacitySchedulerITCase#perJobYarnClusterWithParallelism failed due to NPE