In a single zone(us-east-2a) cluster, modify controlplanemachineset to use OnDelete strategy and add another two zones: us-east-2b and us-east-2c,  the backend subnets for us-east-2b and us-east-2c aren't actually configured. If delete one master machine, even if the machine is up to date and not in need of replacement, a new master will be created in us-east-2a, the old master will be deleted.  

4.12.0-0.nightly-2022-09-12-152748

always

1. Setup a single zone cluster
2. Modify controlplanemachineset to use OnDelete strategy and add another two zones: us-east-2b and us-east-2c 3
$ oc edit controlplanemachineset cluster 
  strategy:
    type: OnDelete
  template:
    machineType: machines_v1beta1_machine_openshift_io
    machines_v1beta1_machine_openshift_io:       
      failureDomains:
        platform: AWS
        aws:
        - placement:
            availabilityZone: us-east-2a
          subnet:
            type: Filters
            filters:
            - name: tag:Name
              values:
              - zhsunaws991-9n7r7-private-us-east-2a
        - placement:
            availabilityZone: us-east-2b
          subnet:
            type: Filters
            filters:
            - name: tag:Name
              values:
              - zhsunaws991-9n7r7-private-us-east-2b
        - placement:
            availabilityZone: us-east-2c
          subnet:
            type: Filters
            filters:
            - name: tag:Name
              values:
              - zhsunaws991-9n7r7-private-us-east-2c  
3. Delete a master machine, the machine is up to date and not in need of replacement 

A new master will be created in us-east-2a instead of us-east-2b and us-east-2c, the old master will be deleted. 

$ oc get machine                         
NAME                                        PHASE      TYPE         REGION      ZONE         AGE
zhsunaws991-9n7r7-master-0                  Deleting   m6i.xlarge   us-east-2   us-east-2a   71m
zhsunaws991-9n7r7-master-1                  Running    m6i.xlarge   us-east-2   us-east-2a   71m
zhsunaws991-9n7r7-master-2                  Running    m6i.xlarge   us-east-2   us-east-2a   71m
zhsunaws991-9n7r7-master-9cwsg-0            Running    m6i.xlarge   us-east-2   us-east-2a   4m43s
zhsunaws991-9n7r7-worker-us-east-2a-cgrcf   Running    m6i.xlarge   us-east-2   us-east-2a   68m
zhsunaws991-9n7r7-worker-us-east-2a-jslhj   Running    m6i.xlarge   us-east-2   us-east-2a   68m
zhsunaws991-9n7r7-worker-us-east-2a-xgh8l   Running    m6i.xlarge   us-east-2   us-east-2a   68m

$ oc get machine               
NAME                                        PHASE     TYPE         REGION      ZONE         AGE
zhsunaws991-9n7r7-master-1                  Running   m6i.xlarge   us-east-2   us-east-2a   94m
zhsunaws991-9n7r7-master-2                  Running   m6i.xlarge   us-east-2   us-east-2a   94m
zhsunaws991-9n7r7-master-9cwsg-0            Running   m6i.xlarge   us-east-2   us-east-2a   27m
zhsunaws991-9n7r7-worker-us-east-2a-cgrcf   Running   m6i.xlarge   us-east-2   us-east-2a   91m
zhsunaws991-9n7r7-worker-us-east-2a-jslhj   Running   m6i.xlarge   us-east-2   us-east-2a   91m
zhsunaws991-9n7r7-worker-us-east-2a-xgh8l   Running   m6i.xlarge   us-east-2   us-east-2a   91m

The current machine is up to date and not in need of replacement. The other two failure domains it will say need replacement and then we would expect those to fail if the backend infrastructure isn't there

https://issues.redhat.com/browse/OCPCLOUD-1503?focusedCommentId=20945295&page=com.atlassian.jira.plugin.system.issuetabpanels%3Acomment-tabpanel#comment-20945295