I have an SSD that is of questionable quality (as featured in this topic).
What happened was that this was an SSD that was bought as the primary drive in my YunoHost home server. After about half a year SMART reported it having received about 1.7 PiB of writes on a 1 TiB drive, therefore it being way over its write count. That drive (Samsung Evo Pro 990) is known to have[1] a firmware bug that reports(?) too much writing. So I assume – but cannot verify – that this is what happened. In any case it will be very hard to predict when it will actually fail.
So, now I am wondering if there is a way to still make use of that Schrödinger’s drive …
What I had in mind – without having actual experience with something like this – was to use it for some cache or to run stuff of it that is also on more reliable HDD (HDDs are in Btrfs RAID-1), so if the SSD fails, it falls back (hopefully automatically) to the HDD without much loss.
Does anyone have any experience with something similar already?