Picking is the scheduler’s final selection phase. After feasibility narrows the workers and scoring assigns values, picking chooses the best candidate according to the policy: either the highest or lowest score. section 1.6 section 10.11
A scheduling decision proceeds through three responsibilities:
- Feasibility identifies workers that can accept the task. section 1.6
- Scoring gives each feasible worker a value based on the scheduling policy. section 1.6 section 10.11
- Picking compares those values and returns the preferred worker or node. section 1.6
Choosing among feasible workers
Suppose three workers pass feasibility. A policy scores them using factors such as resource availability. Picking returns the worker with the highest score, or the lowest when the policy defines lower values as better. section 1.6 section 10.11
Figure 1.9 Mental model for Cube. It has a manager, a worker, and a scheduler, and users (i.e., you) will interact with it via a command line.
Use the mental model to locate picking inside the scheduler and distinguish it from the manager and worker components. section 1.6
