[MAPREDUCE-6776] yarn.app.mapreduce.client.job.max-retries should have a more useful default

• Type: Improvement
• Status: Resolved
• Priority: Major
• Resolution: Fixed
• Affects Version/s: 2.8.0
• Fix Version/s: 2.9.0, 3.0.0-alpha2
• Component/s: client
• Hadoop Flags: Incompatible change, Reviewed
• Release Note: The default value of yarn.app.mapreduce.client.job.max-retries has been changed from 0 to 3. This will help protect clients from failures that are transient. True failures may take slightly longer now due to the retries.
• Assignee: Miklos Szegedi
• Reporter: Daniel Templeton
• Created: 12/Sep/16 20:39
• Updated: 07/Oct/16 22:00
• Resolved: 07/Oct/16 21:52

Description

The default is 0, so any communication failure results in a client failure. Oozie doesn't like that. If the RM is failing over and Oozie gets a communication failure, it assumes the target job has failed. I propose raising the default to something modest like 3 or 5. The default retry interval is 2s.