[HADOOP-2051] JobTracker's TaskCommitQueue is vulnerable to non-IOExceptions

• Type: Bug
• Status: Closed
• Priority: Blocker
• Resolution: Fixed
• Affects Version/s: 0.15.0
• Fix Version/s: 0.15.0
• Assignee: Arun Murthy
• Reporter: Arun Murthy
• Created: 13/Oct/07 09:20
• Updated: 08/Jul/09 16:52
• Resolved: 17/Oct/07 04:39

Description

The JobTracker#TaskCommitQueue#run method only handles IOException}}s. Christian Kunz ran into a scenario where a job was stuck with all tasks in {{COMMIT_PENDING state and the stack traces showed that the "Task Commit Thread" wasn't even around.

The work-around is to model TaskCommitQueue#run along the lines of other long-running threads in the JobTracer (ExpireLaunchingTasks, ExpireTrackers etc.) to catch, log and ignore any Exception in a loop.