This is an interesting time to set up a new node. As you’ve seen there’s already a Badger option to help streamline disk access… and now a new Hashstore option is being tested that should do even better. The idea is to batch more random writes into sequential writes so regular HDDs deal with it better. But those new features aren’t well-tested… yet.
If it was me… I’d set up a completely vanilla new node on ext4… and just let it vet and work towards clearing the nine-month withholding period. Maybe place the DBs on SSD (which is independent, and can be done at any time). But just kinda wait to see if Hashstore becomes the new default option… and then migrate to it.
Or… since it’s new and doesn’t really have data yet… you could yolo it and enable hashstore from the start. You’d be trying something new and experimental… and would probably be fine 
(either way, I’m not sure if I’d set up LVM-Cache or Bcache today, with node internals changing so much)