Openshift nodes with multipath devices require having just one partition labeled as "root"

• https://docs.redhat.com/en/documentation/openshift_container_platform/4.18/html/machine_configuration/machine-configs-configure#rhcos-enabling-multipath-day-2_machine-configs-configure

If Openshift-Virtualization (and HCP) runs over those nodes, with a guest VM running there, and which datadisk (presented as PersistentVolume) is also labeled with the partition label "root" in case the node comes down in a non-graceful way, there could be an overlap of the labels preventing the host node from booting up (as volume won't be never released as explained here https://docs.redhat.com/es/documentation/openshift_container_platform/4.18/html/storage/ephemeral-storage-csi-vol-detach-non-graceful-shutdown )

Steps for reproducing:

1. Schedule a RHEL/RHCOS VM to run on a worker node with its root disk provisioned via FC-multipathed disk.

2. Simulate non-gracefull shutdown on that node. VM datadisk (PV) should remain attached to that node.

3. Try to boot OCP worker node host again

Expected results:

Openshift node should provide a more reliable way to boot from its own root FC-mpathed disk than a generic label that could be used by other devices attached to the node.