A modern multicore x64 with half that RAM would run any workload that cluster could do… faster (carved into a handful of containers or VMs if you needed isolation).

It’s a cool project, for fun! It would be a horrible way to do any useful work though. Just the thought of wiring it all up makes me shudder…

Computers are so fast these days you can have entire VMs in the background and not even notice them. Since Win10/11 come with HyperV and WSL you don’t need to buy any separate hardware to start playing with homelba/Linux stuff.

I’m following this thread more to hear what people would do with unlimited-money setups. Even with limited funds… modern systems are extremely capable: to the point there’s nothing I want to do in a homelab that I can’t already do.

Look at the new E18 pSLC drives on DigitalSpaceport. They’re being bought up by the Chia and ZFS crowds because of high sustained writes and write endurance. They come in 320GB and 640GB versions (essentially 1TB and 2TB models, with pSLC firmware).

From what I understand the first batch is almost bought out, but eventually they have enough of them to be selling through Amazon as well.

But if this is a boot drive… you don’t need anything special. You’re going to use like 1-2% of its rated TBW per year: so any warranty would time out well before you hit endurance limits. Buy a midrange consumer 2-4TB NVMe on Black Friday and be done with it.