[MAPREDUCE-5400] MRAppMaster throws InvalidStateTransitonException: Invalid event: JOB_TASK_COMPLETED at SUCCEEDED for JobImpl

• Type: Sub-task
• Status: Resolved
• Priority: Minor
• Resolution: Cannot Reproduce
• Affects Version/s: 2.0.5-alpha
• Component/s: applicationmaster
• Target Version/s: 2.8.0
• Assignee: Devaraj Kavali
• Reporter: J.Andreina
• Created: 15/Jul/13 13:44
• Updated: 23/Dec/16 20:26
• Resolved: 11/Feb/16 18:49

Description

Step 1: Install cluster with HDFS , MR
Step 2: Execute a job
Step 3: Issue a kill task attempt for which the task has got completed.

Rex@HOST-10-18-91-55:~/NodeAgentTmpDir/installations/hadoop-2.0.5.tar/hadoop-2.0.5/bin> ./mapred job -kill-task attempt_1373875322959_0032_m_000000_0
No GC_PROFILE is given. Defaults to medium.
13/07/15 14:46:32 INFO service.AbstractService: Service:org.apache.hadoop.yarn.client.YarnClientImpl is inited.
13/07/15 14:46:32 INFO proxy.ResourceManagerProxies: HA Proxy Creation with xface : interface org.apache.hadoop.yarn.api.ClientRMProtocol
13/07/15 14:46:33 INFO service.AbstractService: Service:org.apache.hadoop.yarn.client.YarnClientImpl is started.
Killed task attempt_1373875322959_0032_m_000000_0

Observation:
===========
1. task state has been transitioned from SUCCEEDED to SCHEDULED
2. For a Succeeded attempt , when client issues Kill , then the client is notified as killed for a succeeded attempt.
3. Launched second task_attempt which is succeeded and then killed later on client request.
4. Even after the job state transitioned from SUCCEEDED to ERROR , on UI the state is succeeded

Issue :
=====
1. Client has been notified that the atttempt is killed , but acutually the attempt is succeeded and the same is displayed in JHS UI.
2. At App master InvalidStateTransitonException is thrown .
3. At client side and JHS job has exited with state Finished/succeeded ,At RM side the state is Finished/Failed.

AM Logs:
========
2013-07-15 14:46:25,461 INFO [AsyncDispatcher event handler] org.apache.hadoop.mapreduce.v2.app.job.impl.TaskAttemptImpl: attempt_1373875322959_0032_m_000000_0 TaskAttempt Transitioned from RUNNING to SUCCEEDED
2013-07-15 14:46:25,468 INFO [AsyncDispatcher event handler] org.apache.hadoop.mapreduce.v2.app.job.impl.TaskImpl: Task succeeded with attempt attempt_1373875322959_0032_m_000000_0
2013-07-15 14:46:25,470 INFO [AsyncDispatcher event handler] org.apache.hadoop.mapreduce.v2.app.job.impl.TaskImpl: task_1373875322959_0032_m_000000 Task Transitioned from RUNNING to SUCCEEDED
2013-07-15 14:46:33,810 INFO [AsyncDispatcher event handler] org.apache.hadoop.mapreduce.v2.app.job.impl.TaskImpl: task_1373875322959_0032_m_000000 Task Transitioned from SUCCEEDED to SCHEDULED
2013-07-15 14:46:37,344 INFO [AsyncDispatcher event handler] org.apache.hadoop.mapreduce.v2.app.job.impl.TaskImpl: Task succeeded with attempt attempt_1373875322959_0032_m_000000_1
2013-07-15 14:46:37,344 INFO [AsyncDispatcher event handler] org.apache.hadoop.mapreduce.v2.app.job.impl.TaskImpl: task_1373875322959_0032_m_000000 Task Transitioned from RUNNING to SUCCEEDED
2013-07-15 14:46:37,345 ERROR [AsyncDispatcher event handler] org.apache.hadoop.mapreduce.v2.app.job.impl.JobImpl: Can't handle this event at current state
org.apache.hadoop.yarn.state.InvalidStateTransitonException: Invalid event: JOB_TASK_COMPLETED at SUCCEEDED
at org.apache.hadoop.yarn.state.StateMachineFactory.doTransition(StateMachineFactory.java:302)
at org.apache.hadoop.yarn.state.StateMachineFactory.access$300(StateMachineFactory.java:43)
at org.apache.hadoop.yarn.state.StateMachineFactory$InternalStateMachine.doTransition(StateMachineFactory.java:445)
at org.apache.hadoop.mapreduce.v2.app.job.impl.JobImpl.handle(JobImpl.java:866)
at org.apache.hadoop.mapreduce.v2.app.job.impl.JobImpl.handle(JobImpl.java:128)
at org.apache.hadoop.mapreduce.v2.app.MRAppMaster$JobEventDispatcher.handle(MRAppMaster.java:1095)
at org.apache.hadoop.mapreduce.v2.app.MRAppMaster$JobEventDispatcher.handle(MRAppMaster.java:1091)
at org.apache.hadoop.yarn.event.AsyncDispatcher.dispatch(AsyncDispatcher.java:130)
at org.apache.hadoop.yarn.event.AsyncDispatcher$1.run(AsyncDispatcher.java:77)
at java.lang.Thread.run(Thread.java:662)

Issue Links

• is related to: MAPREDUCE-6826 Job fails with InvalidStateTransitonException: Invalid event: JOB_TASK_COMPLETED at SUCCEEDED/COMMITTING