Error performing deploy_step write_image: Command execution failed: Unable to stat device /dev/dm-0p2 after attempting to verify 3 times

Issue

• After OSP 17.1 FFU, scaling out the the new nodes fails at ironic with:

ERROR ironic.conductor.utils [-] Agent returned error for deploy step {'step': 'write_image', 'priority': 80, 'argsinfo': None, 'interface': 'deploy'} on node xxxxxxxx-xxxx-xxxx-xxxx-xxxxxxxxxxxx : Error performing deploy_step write_image: Command execution failed: Unable to stat device /dev/dm-0p2 after attempting to verify 3 times..

• Hardware is Cisco UCS-B200-M5 blades using no local disks and booting on multipathed devices

• When compute host (or controller node or any other role) is booting, it's stuck on a single cursor (not blinking ;)) and nothing happens. If you remove the console from the kernel argument at boot time by pressing "esc" or "shift", it might be on the emergency shell or hung.