[MAPREDUCE-4099] ApplicationMaster may fail to remove staging directory

• Type: Bug
• Status: Closed
• Priority: Critical
• Resolution: Fixed
• Affects Version/s: 0.23.2
• Fix Version/s: 0.23.3, 2.0.2-alpha
• Component/s: mrv2
• Target Version/s: 0.23.3
• Hadoop Flags: Reviewed
• Assignee: Jason Darrell Lowe
• Reporter: Jason Darrell Lowe
• Created: 03/Apr/12 14:00
• Updated: 02/Apr/15 00:42
• Resolved: 11/Apr/12 17:12

Description

When the ApplicationMaster shuts down it's supposed to remove the staging directory, assuming properties weren't set to override this behavior. During shutdown the AM tells the ResourceManager that it has finished before it cleans up the staging directory. However upon hearing the AM has finished, the RM turns right around and kills the AM container. If the AM is too slow, the AM will be killed before the staging directory is removed.

We're seeing the AM lose this race fairly consistently on our clusters, and the lack of staging directory cleanup quickly leads to filesystem quota issues for some users.

Issue Links

• relates to: YARN-2261 YARN should have a way to run post-application cleanup