[SPARK-16956] Make ApplicationState.MAX_NUM_RETRY configurable

• Type: New Feature
• Status: Resolved
• Priority: Major
• Resolution: Fixed
• Fix Version/s: 1.6.3, 2.0.1, 2.1.0
• Component/s: Deploy
• Target Version/s: 1.6.3, 2.0.1
• Assignee: Josh Rosen
• Reporter: Josh Rosen
• Created: 08/Aug/16 19:27
• Updated: 12/Sep/16 22:36
• Resolved: 09/Aug/16 18:25

Description

The ApplicationState.MAX_NUM_RETRY setting, which controls the maximum number of back-to-back executor failures that the standalone cluster manager will tolerate before removing a faulty application, is currently a hardcoded constant (10), but there are use-cases for making it configurable (TBD in my PR). We should add a new configuration key to let users customize this.

Issue Links

• duplicates: SPARK-2424 ApplicationState.MAX_NUM_RETRY should be configurable
• links to: [Github] Pull Request #14544 (JoshRosen)