That's not that easy to say since we don't know anything about your cluster.
If you have no temp space available and don't want to risk anything, maybe get another machine that has enough space, export the dir containing the ML via NFS, mount it on the Stand-By node and install it. I would stop HA on that node graceful - it could be that a reboot is needed after updating which is stated in the log of the update at the end, usually.
If you see your oraclevg always, we can't answer. If you have one active node and one passive Stand-By node, it knows the oraclevg, but can't vary it online since it's vary on on your active node.