you are totally right, personally i’m done with trying to store even semi important data on single drives, it does degrade over time, after learning the nitty gritty on how data storage works, i would say today i wouldn’t trust much else aside from zfs with atleast 1 drive redundancy… if i could spare more i would, but for now that’s something i’m working towards.
5 with 1 redundancy doesn’t seem totally unreasonable for my use case, but i would perfer 10 with 2, but i’ve been told that also comes with issues when trying to resilver a degraded zfs array.
20% capacity seems like a sensible loss to get 1 drive redundancy and i got 12 bays in my server.
so i can do 2 vdevs of 5 drives with 1 redundant, which should allow me to migrate data from one vdev to another and thus upgrade without backing myself into a corner and ofc then i also have 2 spare bays for stuff like hot spares, or simply to add something like adding a 12-16tb drive which could store my current vdev of 5x3tb drives.
so yeah i get what your are saying…
storing data reliablly is quite tricky, especially if one wants to keep functionallity, performance, upgradeability and costs in a balanced range.
Thinking of getting a 36bay shelf next…
HALP! i’m becoming a professional datahorder… xD