[YARN-8118] Better utilize gracefully decommissioning node managers

• Type: Sub-task
• Status: Open
• Priority: Major
• Resolution: Unresolved
• Affects Version/s: 2.8.2
• Component/s: yarn

Environment

• Google Compute Engine (Dataproc)
• Java 8
• Hadoop 2.8.2 using client-mode graceful decommissioning

• Target Version/s: 2.10.3
• Assignee: Karthik Palaniappan
• Reporter: Karthik Palaniappan
• Created: 04/Apr/18 23:11
• Updated: 24/May/22 08:03

Description

Proposal design doc with background + details (please comment directly on doc): https://docs.google.com/document/d/1hF2Bod_m7rPgSXlunbWGn1cYi3-L61KvQhPlY9Jk9Hk/edit#heading=h.ab4ufqsj47b7

tl;dr Right now, DECOMMISSIONING nodes must wait for in-progress applications to complete before shutting down, but they cannot run new containers from those in-progress applications. This is wasteful, particularly in environments where you are billed by resource usage (e.g. EC2).

Proposal: YARN should schedule containers from in-progress applications on DECOMMISSIONING nodes, but should still avoid scheduling containers from new applications. That will make in-progress applications complete faster and let nodes decommission faster. Overall, this should be cheaper.

I have a working patch without unit tests that's surprisingly just a few real lines of code (patch 001). If folks are happy with the proposal, I'll write unit tests and also write a patch targeted at trunk.