-
Type:
Improvement
-
Status: Closed
-
Priority:
Major
-
Resolution: Fixed
-
Affects Version/s: 2.0.4-alpha
-
Fix Version/s: 2.1.0-beta
-
Component/s: scheduler
-
Labels:None
-
Hadoop Flags:Reviewed
Currently, at a regular interval, the fair scheduler computes a fair memory share for each queue and application inside it. This fair share is not used for scheduling decisions, but is displayed in the web UI, exposed as a metric, and used for preemption decisions.
With DRF and multi-resource scheduling, assigning a memory share as the fair share metric to every queue no longer makes sense. It's not obvious what the replacement should be, but probably something like fractional fairness within a queue, or distance from an ideal cluster state.
- depends upon
-
YARN-326 Add multi-resource scheduling to the fair scheduler
-
- Closed
-