Scheduling

• What is Scheduling?

  • Process of assigning Pods to Nodes in a Kubernetes cluster.

• Kubernetes Scheduler:

  • The component responsible for deciding which Node a Pod should run on.
  • Works based on the available resources and the Pod's resource requirements.

• Node Selection:

  • Factors considered by the scheduler:
    • CPU, memory resources.
    • Node affinity.
    • Taints and tolerations.
    • Pod affinity and anti-affinity.
    • Resource requests and limits.

• Scheduling Policies:

  • Default scheduling behavior.
  • Priority scheduling (priority classes).
  • Custom scheduling policies (e.g., using kube-scheduler with custom rules).

• Affinity and Anti-Affinity:

  • Node Affinity: Constraints that allow or prevent Pods from being scheduled on specific nodes.
  • Pod Affinity: Ensures that Pods are scheduled on nodes where other specific Pods are running.
  • Pod Anti-Affinity: Ensures Pods are not scheduled on the same node as other specified Pods.

• Taints and Tolerations:

  • Taints: Applied to Nodes to prevent Pods from being scheduled unless they tolerate the taint.
  • Tolerations: Allow Pods to be scheduled on Nodes with specific taints.

• Resource Requests and Limits:

  • Requests: Minimum amount of resources a Pod needs (scheduler uses this to place Pods).
  • Limits: Maximum amount of resources a Pod can use (enforced during execution).

• Preemption:

  • When a Pod with higher priority displaces lower-priority Pods to be scheduled.

• Scheduler Extenders:

  • Custom schedulers or extender mechanisms to modify default scheduling decisions.

• Pod Disruption Budgets (PDB):

  • Used to manage voluntary disruptions (like upgrades) to ensure availability during disruptions.