SQL Server – Cluster Failover When Running Out of RAM

Welcome to StackExchange.

The SQL database will only need to be installed on 1 node, as the nodes share the storage already since SQL Server is 'cluster aware'. However if you use something that is not 'cluster aware' such as SSIS, then packages will have to be installed on both nodes.

The reason you need to do 2 installations is because the nodes do not share installation files found on the local disk of the machine, but anything you do on shared storage and within SQL Server will be retained when failing over. It's 1 database engine and 2 nodes.

Off topic, but a thought might be to have scripts that copy backup files and such so I make my 'backups' drive a cluster resource, and point all of my scripts to the virtual cluster address so even in case of a failover, they work.

If by AlwaysOn you mean "Availability Groups" (which is typically the case), then you can most definitely use your failover cluster instance as a replica.

For this to be an Availability Groups design that is actually useful in the sense of AGs, then you would most likely want another server to host another replica in the Availability Group (for whatever it is that you're looking for: HA, DR, reporting, etc.). All of the servers that are going to be participating in the Availability Group need to reside in the same WSFC.

There is no point in having AO when only 1 instance of SQL is running at a given time

The replicas in an Availability Group are individual SQL Server instances (whether it be a FCI or standalone instance), so each of the replicas are "running" in a normal condition.

what would I need to set up AO?

Technically speaking? You would just need another instance of SQL Server to have as an additional replica for the AG.

Realistically speaking? We can't answer that right now with the current state of the question. We'd need to know exactly what you want your Availability Group for.