If you can make your dataset fit on a single 24tb disk (or less if you want cheaper disks), that simplifies the setup. I’d use a set of 3 copies: * One main di

If you can make your dataset fit on a single 24tb disk (or less if you want cheaper disks), that simplifies the setup. I’d use a set of 3 copies: * One main disk that you use for collecting and organizing all of the data * One backup disk in the system with the main disk * One backup disk kept off-site in a safe deposit box or similar Every few months, swap the onsite and offsite backup disks to keep the offsite one fresh. When you do that, verify the integrity of the data on the disk that just came back. Automate verifying the integrity of the main and onsite backup monthly. My preference for filesystem would be single-disk ZFS “arrays”. That allows you to run a scrub to verify integrity easily. For copying the data, use either zfs send or rsync. zfs send copies the filesystem directly so it preserves everything about the file (for example sparse files) but rsync is less complex. For getting the data from your various disks, I tend to use dd to create a raw image of the disk, then use other tools like losetup to mount the image as a loop device. Then if there’s data that I want more convenient access to, I’ll rsync it out of the mounted FS. That way, you’ll never lose data from a misconfigured/faulty copy since you can always do it again from the raw disk image.