I’ve decided it’s simpler to just not bother. Also makes “lsblk” output easier to read.
I just do this:
mkfs.ext4 -b 4096 -m 0 /dev/sdc
Curious if anyone else is in this camp? Or do others run cfdisk / fdisk to put the partition table in first?
I use the whole disk too. ext4 for the final plot drive, xfs for the temp file drive.
I have been using xfs for both temp and plot drives without any partition tables!
Unless needing separate spaces to avoid things filling up etc I don’t use partitions except on a couple of disks that have root MD’d across. Has worked fine for thousands of disks for decades, almost all XFS.
Even when I sometimes needed swap for an experiment that could otherwise do OOM badness, I just swapon’d to files on fs. It’s slower than swap to raw partition but if you’re really paging that much you’d probably want to fix it for most app workloads nowadays…