Are these physical servers or virtual machines? Also are they production servers?
Since you only have a two node cluster, the recommended way would be to create an already upgraded third server and add it to your existing cluster. With the third server added you can rebalance your server and remove one of the smaller nodes repeating this process as much as you need. Then if you only wanted to keep two nodes you can gracefully remove the extra node and perform a final rebalance.
By working with a third node in this process, your server won’t experience any extra stress when you remove a node.
Because you are only doing an upgrade you should not initiate failover.
Does this answer your question?