Kube Scheduler selects the node where pod will be placed. It does not create or deploy the pod, it only determines which node is best fir for the pod according to its specification. Once a POD creation request is placed, it has its own memory and CPU requirement, so kubernetes cluster must have process to determines which node can fit the specification for the POD. That is where kube-scheduler performs the right job. Kube-scheduler performs this in a 2 steps process - filter node and rank node. Filter node process eliminated the node which are not suitable for the resource requirement for the pod. Rank node process assigns rank to all the remaining node and the highest rank node is selected.Highest rank is assigned to node which has most number of memory and cpu after assigning the pod resources. You can define custom scheduler algorithm as well. To install scheduler using admin too, it keeps the manifest file at - /etc/kubernetes/manifests/kube-scheduler.yaml and installing it using the kubernetes release binaries, it will be installed as service. To view the running process unser master node, you can do it using ps - aux | grep kube-scheduler
Next