Steps to reproduce the behavior:
Client is doing a minor update 16.1 to 16.2.
Undercloud is successfully updated to 16.2
While doing 'update prepare' command, 80% of the compute and controller nodes were wiped out.
OS was wiped, config lost, containers gone, instances gone.
Except for 1 controller and 3 compute nodes (17 instances in total).

Expected behavior
This should not have happened.

Bug impact
Everything but 17 instances are down.
Control plane is down.

We need help to restore the environment and understand how it happened in the first place.

Right now we have a mysqldump from before the update-prepare and the current state.
Both files are attached to the case.