Thank you very much for answering this. I am running between things I have to do today and fixing this computer is a lower priority until about 5PM local for me…
“Did you assemble the array with mdadm before starting Anaconda? If yes, then you should be able to choose the /dev/md* device as installation target.”
Yes, it was done under SLED. But I do not see how to do this when doing an initial install of Fedora.
" Well, you can’t have a homeless user. What exactly did you change in /etc/fstab? By the way, are you aware of the Discoverable Partition Specification "
I had no idea (the Discoverable Partition Specification) that this existed. And under OPENSuse (in days gone by), one could specify a user ID but tell the system to NOT build the /home entry (because it already existed). In this way one could build everything but the /home because it was already in existence from a prior Linux install. So all one needed to do at that point is define the users, set their passwords, and they were ready to run because all their data was already in place.
fstab: I changed that in SLED to get it to pick it back up (SLED got damaged somehow in the install – seems they were having problems with repositories last night – seems I hit this when doing maint on weekends). And I am answering using an older SLED 12 image (lost my SLED 15 DVD, just reburned it).
Now back to the mount point: Anaconda does not (that I could find) allow me to tell it to pickup the RAID1 partitions (even though it recognized them and knew which was “primary”…) unless I told it to reformat. This, in my opinion, is a design flaw. If one can effectively do an upgrade from one Linux to another…
The possible catch in this is the ID numbering, which I think Suse starts with 1000… The other is going to be the group definitions, but that is a different problem that must be resolved by the “accidental administrator” (me). I have been a user of LInux desktops since about 2002, but I am not a Linux power user. Just know enough to be ugly dangerous (thankfully my day job has to do with Mainframes where I have been a systems programmer (super admin), and internals software developer. So I have some idea of what an O/S needs for I/O definitions since I work with z/OS, zVM, and zVSE).