OpenShift OKD – Cluster API Scaling Specific Nodes

Scaling down specific nodes in OpenShift / OKD can be achieved using the cluster-api-operator and a specific node annotation.
https://github.com/openshift/machine-api-operator/blob/release-4.6/pkg/controller/machineset/delete_policy.go#L34

This is a way to ensure predetermined nodes are removed and can be useful, particularly in the case of Baremetal nodes if you need to remove one specifically for whatever reason.

I’m using OKD deployed on OpenStack in this video.

We discuss deletionPolicy, what alternatives exist, and how to select predetermined nodes for deletion:

https://docs.okd.io/latest/machine_management/manually-scaling-machineset.html#machineset-delete-policy_manually-scaling-machineset

Default Policy is Random:
https://github.com/openshift/machine-api-operator/blob/release-4.6/pkg/controller/machineset/delete_policy.go#L80-L91

source by TripleWho?

redhat openstack