[YARN-3136] getTransferredContainers can be a bottleneck during AM registration

• Type: Sub-task
• Status: Closed
• Priority: Major
• Resolution: Fixed
• Affects Version/s: 2.6.0
• Fix Version/s: 2.8.0, 2.7.2, 3.0.0-alpha1
• Component/s: scheduler
• Labels: 2.7.2-candidate
• Target Version/s: 2.8.0
• Hadoop Flags: Reviewed
• Assignee: Sunil G
• Reporter: Jason Darrell Lowe
• Created: 04/Feb/15 18:58
• Updated: 02/Jan/18 18:09
• Resolved: 18/Apr/15 19:49

Description

While examining RM stack traces on a busy cluster I noticed a pattern of AMs stuck waiting for the scheduler lock trying to call getTransferredContainers. The scheduler lock is highly contended, especially on a large cluster with many nodes heartbeating, and it would be nice if we could find a way to eliminate the need to grab this lock during this call. We've already done similar work during AM allocate calls to make sure they don't needlessly grab the scheduler lock, and it would be good to do so here as well, if possible.