30. Explanation of core components of Kubernetes - Scheduler

This chapter explains the knowledge points

• 1. Scheduler overview
• 2. Scheduler principle analysis
• 3. The two-stage scheduling process of the old version of Scheduler
• 4. Scheduler Framework

---

1. Scheduler overview

Kubernetes Scheduler is a core component in the Kubernetes cluster, which is used to schedule Pods to run on suitable Nodes . In Kubernetes, Pod is the smallest schedulable unit, and Node is the physical or virtual machine running Pod . The scheduler is responsible for deciding which Node each Pod should run on to maximize the use of cluster resources and satisfy user-defined scheduling policies.

Kubernetes Scheduler implements Pod scheduling through scheduling algorithms and policies. The scheduling algorithm can be priority scheduling, fair scheduling, weight scheduling, etc., and the strategy can be node affinity, Pod affinity, taint tolerance, etc. When scheduling, the Scheduler will consider a series of factors, such as node resource utilization, Pod resource requirements, node affinity, Pod affinity, taint tolerance, etc., to determine the best scheduling scheme.

In addition to the default scheduler, Kubernetes also allows users to use custom schedulers to meet specific scheduling needs, such as scheduling Pods to GPU nodes or scheduling Pods to specific geographic locations.

Kubernetes Scheduler provides services through a process called kube-scheduler, which runs on the Master. It belongs to static pod. </