How to recover the Virtual Machine instance which is in the error state after a failed resize attempt

Issue

• The Virtual Machine (VM) instance status transitions to the error state when it is resized to a different flavor than the original one. While resizing, the instance was attempted to be moved to a different compute node than the source one. But the destination compute node had some missing Fibre Channel device and failed to cater to the request of accommodating the resized VM instance. At this stage, the rollback resize attempt of the affected VM instance to the original compute node fails too. The Cinder volume logs (/var/log/containers/cinder/cinder-volume.log) show the following error.
  <date_and_timestamp> <event_number> ERROR cinder.volume.driver [req-<id> <id> <id> - default default] FibreChannelDriver validate_connector failed. No 'wwpns'. Make sure HBA state is Online.