Interesting question. I just checked, and the oldest plot I’m actively farming is from July 8, 2021. I have retired a bunch of plots on a drive that started failing, so I might have been farming as much as a week earlier.
I have had good luck with drives that I have hand-picked. When I left my previous place of employ, a pair of Linux servers that I had specified individual parts for was still running 18 1/2 years after they were built, the Hitachi Ultrastar SCSI drives not having a single failure in all that time.
I just retired the oldest drives I have at home, four 2TB Samsung Spinpoint backing my MythTV server and NAS. While my farming mostly used an assortment of USB-attached drives I had laying about, some of the plots were on the Spinpoint NAS. One of the four drives started returning read errors a few months ago, triggering their retirement. Purely coincidentally, that was just a couple of months after I put together a new NAS using 14TB Toshiba drives for capacity reasons. In any case, those Samsung Spinpoints were running 14 years 10 months when pulled from service.
On the flip side, I avoid Western Digital like the plague. Those computers I noted above, running 18 1/2 years, replaced another pair that were outfitted with WD drives. When we think of old Linux establishments that are nearly historic, we think of RedHat and Suse and Debian; but this earlier pair predates all of those! Alas, the WD drives all started failing spectacularly within weeks of each other at about the 2 year mark. I had just enough time to put together the cluster with Hitachi drives before these would no longer initialize after a power cycle. Somebody suggested to me that “Western Digital is a lot better than they used to be all those years ago, give them a try.” I did, and bought a WD portable backup drive, USB connected. This new WD drive lasted 1 hour 30 minutes before it started making “click click” noises, and never powered on again. Take this sad tale “with a grain of salt” as they say.
