[PIG-5327] Check for DAG status before trying to kill

• Type: Bug
• Status: Resolved
• Priority: Major
• Resolution: Fixed
• Fix Version/s: 0.18.0
• Hadoop Flags: Reviewed
• Assignee: Rohini Palaniswamy
• Reporter: Rohini Palaniswamy
• Created: 06/Jan/18 00:56
• Updated: 09/Jan/18 13:26
• Resolved: 09/Jan/18 13:26

Description

The killJob is invokedby

• grunt kill command
• thread interrupt in Main (mostly in case of Oozie)
• HangingJobKiller during shutdown

killJob tries to kill the DAG again even if it was completed. It closes the session but does not remove it from the TezSessionManager session pool. Though both of these are minor issues and have no adverse impact (trying to kill a done DAG would be ignored by Tez AM and TezSessionManager would later check status of the session and remove it from the pool) the fix would avoid the two unnecessary calls to AM.