[YARN-4414] Nodemanager connection errors are retried at multiple levels

• Type: Bug
• Status: Closed
• Priority: Major
• Resolution: Fixed
• Affects Version/s: 2.7.1, 2.6.2
• Fix Version/s: 2.8.0, 2.7.3, 2.6.4, 3.0.0-alpha1
• Component/s: nodemanager
• Target Version/s: 2.7.3, 2.6.4
• Hadoop Flags: Reviewed
• Assignee: Chang Li
• Reporter: Jason Lowe
• Created: 03/Dec/15 17:01
• Updated: 06/Jan/17 01:44
• Resolved: 12/Jan/16 16:02

Description

This is related to YARN-3238. Ran into more scenarios where connection errors are being retried at multiple levels, like NoRouteToHostException. The fix for YARN-3238 was too specific, and I think we need a more general solution to catch a wider array of connection errors that can occur to avoid retrying them both at the RPC layer and at the NM proxy layer.

Issue Links

• relates to: YARN-3238 Connection timeouts to nodemanagers are retried at multiple levels