"I need a HDD with failing SMART data" and other phrases indicative of moderately cursed software development projects
Actually, let's make this into a public request; if you have a HDD with failing SMART attributes of any sort, please run `smartctl --attributes --json` with a decently recent version (the nixpkgs one is definitely good enough) and send me the output!
(I'm specifically interested in how failing metrics are represented, so completely healthy drives are not useful here)
@joepie91 I've got one! I've had a failing disk on my NAS which I keep delaying the replacement for.
https://paste.sr.ht/~zurdo/4b16a31b6e036ad3add40c617f3e35a4bd556efb
That's what I've found today when running the command. Apparently it fully died at some point in the last week, I hoped to find some actual data, but alas, I guess I'm finally replacing it today 
EDIT: Since zfs is fucking dumb I had to plug it again and I got actual data back this time
https://paste.sr.ht/~zurdo/ce227937d0e95f02152a70390c6ee0ca95c91f26
@Zurdo Thanks! This is very helpful, and confirms what I suspected - the when_failed values are different in the JSON output compared to the plaintext output :)
(Now to find someone with a disk that is "failing now" and my data should be complete :p)
@Zurdo Oh, actually, this helped me find the responsible code in the source code: https://github.com/mirror/smartmontools/blob/4591526a9de7dceca255783390236bcdf61778b7/ataprint.cpp#L1285
So my data is already complete now, thanks :)