How to fit as many plots as possible on a 16TB hdd?

There is no such “array.” That’s why it’s called, “Just a bunch of disks” … like a bunch of disks in your PC.

You are confusing JBOD and RAID 0 (Stripe). Raid 0, yes you will lost all your data when one disk fails. JBOD is just separately connected disks and they are not magically joined in any way and you do not get a part file from one transferred to the other when space runs out. It might be possible that some NAS implementation does this but I’m not aware of it. In JBOD one disk files and you only lose that disk - there is also no space lost for parity, because there is no parity and it is not an array of any sort.

Here’s a solution I wrote up for Windows users - use virtual disks and a storage pool to join the last <100gb of space on each drive into one big drive and then put more plots into that. If one of the “crumbs” of this drive gets corrupted, you might lose several plots but no biggie - just replot and keep going.

Actually there is a historic use of jbod. It means the disks are joined together hence Just A Bunch of Disks.
The current use of jbod is where each disk is passed through to the OS individually.
It should be called JD - just disks. There’s no bunch in it anymore

That’s interesting. What is the difference between that historic JBOD and RAID 0 / STRIPE?

If I remember right the disks could be different sizes because they weren’t striped.

You might be thinking about unraid? Which still does that now, but that’s definitely not jbod.