This maintenance resulted in unexpected hardware failures of two host blades in this cluster.
This caused VM HA events to fail over the VMs from the failed blades onto remaining capacity in the cluster.
We are still investigating the reason for the hardware failures during the upgrade, and will bring up a case with the vendor to determine what happened, and prevent it in the future.
Any VM that HA failed over should be rebooted online at this point.
This cluster is stable and running the expected workload at this time.
Other than the unexpected hardware failures, the upgrade has completed successfully.