Kubernetes Classroom Series – 02/Jul/2021

Scheduling Process (Automated Placement)

• Overview

• As soon as pod is create that is not assigned to a node yet, it gets picked by scheduler together with all the available nodes and set of filtering and priority policies.

• In first stage, Scheduler tries to apply the filtering and priority policies by removing all the nodes that do not qualify the Pod’s criteria

• In Second stage, all the qualified nodes get ordered by weight

• In last stage Pod gets assigned to the node (highest weight)

• Node Selctor:

  • This assignment can be done under node selector
  • .spec.nodeSelector is the Pod field and specifies a map of key values pairs that must be present on the node for the node to be eligible to run the Pod
  • To assign labels to nodes Refer Here

    apiVersion: v1
    kind: Pod
    metadata:
    name: nodeselector-demo
    spec:
    containers:
      - name: nodeselector-container
        image: nginx
        ports:
          - containerPort: 80
        livenessProbe:
            httpGet: 
            port: 80
            path: /
            initialDelaySeconds: 30
    nodeSelector:
        disktype: ssd
        os: ubuntu    
    
    

    • Refer Here for the changeset

• Node Affinity:

  • This feature is generalization of the node selector approach.
  • Node Affinity feature greatly expands the types of constraints that be expressed by providing operators
    • In
    • NotIn
    • Exists
    • DoesNotExist
    • Gt
    • Lt
  • Refer Here for the changeset containing node affinity demo