Upgrade may fail if operations such as such as OS reboot, cluster restart, and node stop and shutdown are used during the upgrade
Performing any of the following operations during an upgrade can lead to an upgrade failure and the cluster might go in an inconsistent state. It is recommended that these operations should not be performed until the upgrade completes successfully. (IA-40377)
Logging into maintenance mode and executing OS commands such as reboot and shutdown.
Logging into a node and executing commands such as cluster reboot, shutdown, and stop from the command-line interface.
Executing node maintenance operations from the UI such as shutdown node and start node.
Operations that will disrupt network connections such as unplugging the network cable.
Workaround:
There is no workaround for this issue.