ChipTuner on Nostr: So, for me, it's way more effective to manage one large disk (ssd or hdd) array and ...

So, for me, it's way more effective to manage one large disk (ssd or hdd) array and share it across all cluster nodes, with proper backups, hardware redundancy, and so on, instead of a hyper converged setup. At that point I might as well just dedicate hardware to individual machines, defeating the entire purpose of a cluster. I'm over the days of running individual consumer machines for tasks, not worth it in the long run. I have to make some compromises but getting deals on enterprise gear is where it's at IMO.

And yeah I do intend to move everything back to the slow storage for better backups, but for now it seems to be cooking better.

Around here the sweet spot is 512gb disks.