On Wed 17 Jan 2024 at 15:34:09 (-0500), gene heskett wrote: > On 1/17/24 12:27, Thomas Schmitt wrote: > > David Christensen wrote: > > > I suspect the conflicting serial numbers are causing problems in the > > > kernel, > > > as indicated by the /dev/disk/by-id/* problems. > > > > That's not in the kernel but in udev/systemd's process of creating the > > symbolic links in /dev/disk/by-id/. > > It gets /dev/sd[h-l] and /dev/sd[h-l]1 as kernel generated device files. > > But sd[ij] and also sd[hl] show pair-wise the same serial numbers. > > In case of sd[ij] the outcome is mixed: links to sdi and sdj1 survive. > > In case of sd[hl] we see a less strange outcome: sdh and sdh1, while > > sdl and sdl1 are missing. > > > missing because the original command line did not look at sdl. > I added the l and it showed up. No magic.
What do you mean, it was "missing"? The original command, which I wrote for you, contained a wildcard, so it doesn't miss anything that's there: root@coyote:~# for j in /dev/disk/by-id/* ; do printf '%s\t%s\n' "$(realpath "$j")" "$j" ; done and there was no sdl in the output from that command. In fact, there was no "l" in your post between the "l" in "realpath", above, and the "l" in "like", below: root@coyote:~# but like I wrote, 2 pairs with identical "serial numbers", so the https://lists.debian.org/debian-user/2024/01/msg00658.html shows this, that no sdl was seen under by-id/. > > The open question (at least to me) is whether it's the disks or the > > controllers or the drivers which cause the duplication. > Neither, a typu in the original command. Cheers, David.
