[FLINK-4142] Recovery problem in HA on Hadoop Yarn 2.4.1

• Type: Bug
• Status: Closed
• Priority: Major
• Resolution: Fixed
• Affects Version/s: 1.0.3
• Fix Version/s: 1.1.0
• Component/s: Deployment / YARN
• Assignee: Robert Metzger
• Reporter: Stefan Richter
• Created: 01/Jul/16 14:29
• Updated: 19/Dec/16 12:46
• Resolved: 15/Jul/16 13:01

Description

On Hadoop Yarn 2.4.1, recovery in HA fails in the following scenario:

1) Kill application master, let it recover normally.
2) After that, kill a task manager.

Now, Yarn tries to restart the killed task manager in an endless loop.

Issue Links

• is related to: YARN-2065 AM cannot create new containers after restart-NM token from previous attempt used
• links to: GitHub Pull Request #2255