10.5.8.2. Sample RHOSP failure domain configuration

To prevent downtime for your application due to the failure of a single Red Hat OpenStack Platform (RHOSP) region, you can configure failure domains in the control plane machine set. To use failure domains, you configure appropriate values in the failureDomains section of the ControlPlaneMachineSet custom resource (CR).

The control plane machine set concept of a failure domain is analogous to the existing RHOSP concept of an availability zone. The ControlPlaneMachineSet CR spreads control plane machines across more than one failure domain when possible.

apiVersion: machine.openshift.io/v1
kind: ControlPlaneMachineSet
metadata:
  name: cluster
  namespace: openshift-machine-api
spec:
# ...
  template:
# ...
    machines_v1beta1_machine_openshift_io:
      failureDomains:
        openstack:
        - availabilityZone: nova-az0
          rootVolume:
            availabilityZone: cinder-az0
        - availabilityZone: nova-az1
          rootVolume:
            availabilityZone: cinder-az1
        - availabilityZone: nova-az2
          rootVolume:
            availabilityZone: cinder-az2
        platform: OpenStack
# ...

where: