[YARN-3562] unit tests failures and issues found from findbug from earlier ATS checkins

• Type: Sub-task
• Status: Resolved
• Priority: Minor
• Resolution: Fixed
• Affects Version/s: YARN-2928
• Fix Version/s: 2.9.0, 3.0.0-alpha1
• Component/s: timelineserver
• Target Version/s: YARN-2928
• Hadoop Flags: Reviewed
• Assignee: Naganarasimha G R Assign to me
• Reporter: Sangjin Lee
• Created: 29/Apr/15 18:50
• Updated: 25/Oct/19 20:26
• Resolved: 07/May/15 03:36

Description

Issues reported from MAPREDUCE-6337 :
A bunch of MR unit tests are failing on our branch whenever the mini YARN cluster needs to bring up multiple node managers.

For example, see https://builds.apache.org/job/PreCommit-MAPREDUCE-Build/5472/testReport/org.apache.hadoop.mapred/TestClusterMapReduceTestCase/testMapReduceRestarting/

It is because the NMCollectorService is using a fixed port for the RPC (8048).

Issues reported from YARN-3044 :
Test case failures and tools(FB & CS) issues found :

1. find bugs issue : Comparison of String objects using == or != in ResourceTrackerService.updateAppCollectorsMap
2. find bugs issue : Boxing/unboxing to parse a primitive RMTimelineCollectorManager.postPut. Called method Long.longValue()
   Should call Long.parseLong(String) instead.
3. find bugs issue : DM_DEFAULT_ENCODING Called method new java.io.FileWriter(String, boolean) At FileSystemTimelineWriterImpl.java:[line 86]
4. hadoop.yarn.server.resourcemanager.TestAppManager, hadoop.yarn.server.resourcemanager.rmapp.TestRMAppTransitions, hadoop.yarn.server.resourcemanager.TestClientRMService & hadoop.yarn.server.resourcemanager.logaggregationstatus.TestRMAppLogAggregationStatus, refer https://builds.apache.org/job/PreCommit-YARN-Build/7534/testReport/

Issue Links

• is related to: MAPREDUCE-6337 add a mode to replay MR job history files to the timeline service
• is related to: YARN-3044 [Event producers] Implement RM writing app lifecycle events to ATS
• relates to: YARN-3634 TestMRTimelineEventHandling and TestApplication are broken