[TEZ-1493] Tez examples sometimes fail in cases where AM recovery kicks in

• Type: Bug
• Status: Closed
• Priority: Blocker
• Resolution: Fixed
• Fix Version/s: 0.5.0
• Target Version/s: 0.5.0
• Assignee: Jeff Zhang
• Reporter: Jeff Zhang
• Created: 25/Aug/14 09:42
• Updated: 06/Sep/14 01:35
• Resolved: 27/Aug/14 00:48

Description

14/08/25 17:37:03 INFO client.TezClient: Submitting DAG to YARN, applicationId=application_1408499461970_0053, dagName=WordCount
14/08/25 17:37:03 INFO impl.YarnClientImpl: Submitted application application_1408499461970_0053
14/08/25 17:37:03 INFO client.TezClient: The url to track the Tez AM: http://jzhangMBPr.local:8088/proxy/application_1408499461970_0053/
14/08/25 17:37:03 INFO client.RMProxy: Connecting to ResourceManager at /0.0.0.0:8032
14/08/25 17:37:03 INFO client.AHSProxy: Connecting to Application History server at /0.0.0.0:10200
14/08/25 17:37:03 INFO rpc.DAGClientRPCImpl: Waiting for DAG to start running
14/08/25 17:37:07 INFO rpc.DAGClientRPCImpl: DAG: State: RUNNING Progress: 0% TotalTasks: 2 Succeeded: 0 Running: 0 Failed: 0 Killed: 0
14/08/25 17:37:15 INFO rpc.DAGClientRPCImpl: DAG: State: RUNNING Progress: 50% TotalTasks: 2 Succeeded: 1 Running: 0 Failed: 0 Killed: 0
14/08/25 17:37:17 INFO rpc.DAGClientRPCImpl: DAG completed. FinalState=SUBMITTED
WordCount failed with diagnostics: []

The client side shows that the job is failed, but checking the logs found that the recovery works in server side, and eventually finish the job successfully.