SSDs do implement smart and usually do give an early warning.

I've destroyed a few Samsung 840 Evo drives in a few servers and didn't lose a single byte. Failure condition was a very degraded read speed. It took a few hours to copy the 120GB to a fresh SSD

> I get that SSD don't implement S.M.A.R.T

Despite the S.M.A.R.T data is less predictive on SSDs, they do have it, at least for SATA disks...

NVMe drives have it as well, at least the Samsung ones.

You can read it on Linux by running:

"nvme smart-log /dev/nvmeX"

smartmontools can grab most of that info too, even on Windows.

In which case, even less reason to hate on SSD for home storage in RAID shells.

I have a spun-down RAID farm as longterm. It's worrying me that the unit the disks are in is a SPARC based motherboard now over 10 years old. Those caps don't last forever.

and the magnetized bits on an HDD have an expected livetime of 25 years. Unless read-and-rewritten every so many years