[MAPREDUCE-3031] Job Client goes into infinite loop when we kill AM

• Type: Bug
• Status: Closed
• Priority: Blocker
• Resolution: Fixed
• Affects Version/s: 0.23.0
• Fix Version/s: 0.23.0
• Component/s: mrv2
• Hadoop Flags: Reviewed
• Assignee: Siddharth Seth
• Reporter: Karam Singh
• Created: 19/Sep/11 13:13
• Updated: 15/Nov/11 00:50
• Resolved: 26/Sep/11 17:32

Description

Started a cluster. Submitted a sleep job with around 10000 maps and 1000 reduces.
Killed AM with kill -9 by which time already 7000 thousands maps got completed.

On the RM webUI, Application is stuck in Application.RUNNING state. And JobClient goes into an infinite loop as RM keeps telling the client that the application is running.

Issue Links

• duplicates: MAPREDUCE-3072 NodeManager doesn't recognize kill -9 of AM container
• duplicates: MAPREDUCE-2875 NM does not communicate Container crash to RM