[MAPREDUCE-6230] MR AM does not survive RM restart if RM activated a new AMRM secret key

• Type: Bug
• Status: Closed
• Priority: Blocker
• Resolution: Fixed
• Fix Version/s: 2.7.0, 2.6.1, 3.0.0-alpha1
• Component/s: mr-am
• Labels: 2.6.1-candidate
• Target Version/s: 2.7.0
• Hadoop Flags: Reviewed
• Assignee: Jason Darrell Lowe
• Reporter: Jason Darrell Lowe
• Created: 27/Jan/15 16:13
• Updated: 30/Aug/16 01:19
• Resolved: 28/Jan/15 23:52

Description

A MapReduce AM will fail to reconnect to an RM that performed restart in the following scenario:

1. MapReduce job launched with AMRM token generated from AMRM secret X
2. RM rolls new AMRM secret Y and activates the new key
3. RM performs a work-preserving restart
4. MapReduce job AM now unable to connect to RM with "Invalid AMRMToken" exception

Issue Links

• is related to: MAPREDUCE-6324 Uber jobs fail to update AMRM token when it rolls over
• is related to: YARN-3103 AMRMClientImpl does not update AMRM token properly