To Reproduce Steps to reproduce the behavior:

1. Do FFU 16.2 -> 17.1
2. Run the following command. In this first attempt, "OVNDBs VIP" is removed from pacemaker, but this command fails at subsequent tasks

   $ openstack overcloud upgrade run --yes --stack akb-overcloud --debug --limit allovercloud,undercloud --playbook all 
      :
   ... CHANGED | Remove OVNDBs VIP from pacemaker | ...
      :
   ... FATAL | xxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxx | ...

1. Retry the upgrade, but this time it fails on `Remove OVNDBs VIP from pacemaker` tasks because the OVNDBs VIP has been removed at the first attempt

   $ openstack overcloud upgrade run --yes --stack akb-overcloud --debug --limit allovercloud,undercloud --playbook all 
      :
   ... FATAL | Remove OVNDBs VIP from pacemaker | ... | error={"changed": false, "error": "", "msg": "Failed, to set the resource  to the state delete", "output": "\nUsage: pcs resource delete...\n    delete <resource id|group id|bundle id|clone id>\n        Deletes the resource, group, bundle or clone (and all resources within\n        the group/bundle/clone).\n\n", "rc": 1} 

1. The error occurs at the following tasks. In the second attempt, the OVNDBs VIP has been removed. Therefore, "ovn_vip.output" will be empty at the "Fetch ovn VIP" task. Then "Remove OVNDBs VIP from pacemaker" task fails because empty is passed to the "pacemaker_resource" module.

           - name: Fetch ovn VIP
             shell: |
               pcs constraint list |grep "with ovn-dbs-bundle" |awk '{print $1}'
             register: ovn_vip
              :
           - name: Remove OVNDBs VIP from pacemaker
             when:
              - step|int == 5
              - "{{(ovn_dbs_short_bootstrap_node_name|lower == ansible_facts['hostname']|lower)|bool}}"
             pacemaker_resource:
               resource: "{{ ovn_vip.stdout }}"
               state: delete

Expected behavior

• "openstack overcloud upgrade run" succeeds even on the second attempt.

Bug impact

• This may affect many customers doing the upgrade, and this bug will be a blocker of upgrades

Known workaround

• We've not verified yet, but I think we can avoid this issue by modifying THT manually like the following example

  /usr/share/openstack-tripleo-heat-templates/deployment/ovn/ovn-dbs-cluster-ansible.yaml
  
          - name: Remove OVNDBs VIP from pacemaker
            when:
             - step|int == 5
             - "{{(ovn_dbs_short_bootstrap_node_name|lower == ansible_facts['hostname']|lower)|bool}}"
             - ovn_vip.stdout | length > 0 <==== Add this line
            pacemaker_resource:
              resource: "{{ ovn_vip.stdout }}"
              state: delete 

• And then rerun `openstack overcloud upgrade prepare` and `openstack overcloud upgrade run`