Following this K8s page here on how to use multiple schedulers,and this, all I could see is that the schedulerName
is taking place under the pod's spec , though I don't understand if this is explained particularly to single pods (as Pod is the smallest k8s deployable object), or even when a pod is attached to deployment or other deployable resources. In my case, I have a custom scheduler, & I want it to handle scheduling a Daemonset object. I have tried two options:
a. Put the spec.schedulerName: custom-scheduler
under the daemonset pod's specs.
b. Put the spec.schedulerName: custom-scheduler
under the daemonset specs.
And the results are:
a. The pods of the daemonset were scheduled by the custom-scheduler as can be seen by the pods events.
b. The pods of the daemonset were scheduled by the default scheduler of kube-system.
My question is:
Can the user determine the scheduler for deployable k8s objects other than Pods, for example daemonset/ deployment/replicas? if not, please explain why & how this works internally. Your insights are very appreciated.