[YARN-3842] NMProxy should retry on NMNotYetReadyException

• Type: Bug
• Status: Closed
• Priority: Critical
• Resolution: Fixed
• Affects Version/s: 2.7.0
• Fix Version/s: 2.8.0, 2.7.1, 2.6.4, 3.0.0-alpha1
• Target Version/s: 2.7.1, 2.6.4
• Hadoop Flags: Reviewed
• Assignee: Robert Kanter
• Reporter: Karthik Kambatla
• Created: 16/Jun/15 22:49
• Updated: 06/Jan/17 00:51
• Resolved: 23/Jun/15 00:52

Description

Consider the following scenario:
1. RM assigns a container on node N to an app A.
2. Node N is restarted
3. A tries to launch container on node N.

3 could lead to an NMNotYetReadyException depending on whether NM N has registered with the RM. In MR, this is considered a task attempt failure. A few of these could lead to a task/job failure.

Issue Links

• relates to: YARN-3839 Quit throwing NMNotYetReadyException