Skip to content

KodeCloud CKA Kube Scheduler

Links: 111 KodeCloud Index


It is only responsible for deciding which pod goes on which node, it DOESN'T actually place the pod on the node.

That is the job of the kubelet.

  • If we are setting up the scheduler from scratch then we can download the binary and run it as a service.
  • kubeadm tool deploys the kube-scheduler as a pod in the kube-system namespace on the master node.
    • Location to view the configuration on the master node: cat /etc/kubernetes/manifests/kube-scheduler.yaml
How does a scheduler work in the backend?
  • Every pod has a field called nodeName in spec that by default is not set.
  • We don't specify this field when we create the pod manifest file. k8s adds it automatically.
  • The scheduler goes through all the pods and looks for those that do not have this property.
  • It then runs the scheduling algorithm to find the right node for the pod by setting the nodeName property to the name of the node.
  • It does this by creating a binding object.
  • So if don't have a scheduler then one way of scheduling the pods would be specifying the nodeName
  • A pod remains in a PENDING state unless it is scheduled by the scheduler.

  • If for some reason the default scheduler doesn't meet our needs we can write our own scheduler.

    • When creating a pod or a deployment we can instruct k8s to have pod scheduled by our custom scheduler.

How does the scheduler work?

  • Lets suppose we submit a pod definition file to the kubeapi-server.
  • It asks the scheduler to find the right node for the pod depending on the requirements.
  • The first step in the creation of pods is that they end up in a scheduling queue.
    • This is where the pods wait to be scheduled.
  • Pods are sorted based on the priority defined on the pods in the scheduling queue.
  • To set a priority we must first create a priority class/object and set a priority value.
  • attachments/Pasted image 20221012230631.jpg
  • After the sorting/queuing phase the pods enter the filtering phase where the nodes that cannot run the pod are filtered out.
  • After the filtering phase we have the scoring phase.
    • In this nodes are scored with different weights.
    • On the remaining nodes the scheduler associates a score with each node after reserving the CPU for the pod.
    • attachments/Pasted image 20221012231013.jpg
  • The second node has the higher score and hence it gets picked up.
  • After the scoring phase we have the binding phase in which the pod is bound to the node.
  • All of these processes are achieved using a plugin.
    • attachments/Pasted image 20221012231325.jpg
  • The highly customisable nature of k8s ensures that we can write our own plugin.
    • This is achieved using extension points.
  • At each stage there is an extension point to which a plugin can be plugged to.
    • attachments/Pasted image 20221012231551.jpg
    • Some plugins extend to multiple extension points.

Last updated: 2022-10-12