So today, after “stalking” packman waiting for it to populate the 16.0 portion of their repo, I tried to install via USB. ZOTAC with MSI/AMD/nVidia onboard stuff, etc. 128 GB RAM, two NVMEs.
I got hung up on partitioning in 16 during install (after learning to pass nomodeset). I tried all kinds of options and read the documentation and asked on Reddit for help and still could not install in a satisfactory way. I prefer LUKS+LVM because I run a home system (and all I need to prevent is someone from stealing my tax returns, so nothing serious). LUKS+LVM has been my method since like Slackware 9 (i.e. unencrypted /boot, with root, swap and /home in the LUKS encrypted LVM). There were no RAID options (I was going to try using software RAID1 for the first time) either. I tried and tried and finally told it to dump all my openSUSE 15.6 to install encrypted LUKS using FDE. I watched as selecting FDE encrypted my LVM, and I was cool with that, though I don’t need FDE and I had to dump everything.
What made it worse is I provided the agama installer with my LUKS password at the beginning. It never populated my LVM at all. So I couldn’t choose current partitions.
When I did ctrl+alt+F5 to go into virtual CLI, I checked the partitions to get the number so I could make sure I didn’t delete anything important. ctrl+alt+F7 did not put me back into the installer because it asked for username and password, which was not listed anywhere. So I had to reboot and try again.
Welp, FDE really did mean full disk encryption, as I got the dreaded GRUB2 prompt before the menu. I detest it because it’s a very slow option that provides no real benefits and could be argued to put the system at risk by potentially passing info to Windows. But whatever.
So first problem: install options are not good (they’re almost like Ubuntu, which is annoying because it feels like my intelligence is being insulted). No RAID (see another user’s write-up) either.
At first boot, there was no Windows option in GRUB2 either. Ok, I thought, not a problem, I’ll go into Yast-BootLoader (because while I used Slackware-Current for 10 years, I never was able to figure out how to work with GRUB2). No Yast anywhere to be found even though I told it XFCE desktop with office and creativity software (there was no way to pick specific software either, another complaint).
So I got on WiFi and installed all the Yast2 tools using Zypper. None of them worked. Maybe I got lucky because one of the installs triggered a grub2 update, so the next reboot showed Windows in the GRUB2 menu.
But Yast2 continued to not work. The funny thing is I began using OpenSUSE like 10 years ago because my life changed: college. Now the amazing Yast2 tools which set OpenSUSE apart from all other distributions seem to be broken and my disk options are gone in the installer.
So I went back to 15.6 (and even put in software RAID1 while I was at it). I have long been an advocate of OpenSUSE even while I was an IT instructor. I taught it to my students.
Apparently some developer on Reddit is telling us that agama use case survey revealed that “no one” uses LUKS+LVM like I do. This means, as a logical extension of what this developer said, “no one” uses RAID 1 (see other bro’s post about how to install with RAID1). And this developer seems to be actively telling people that it’s “too bad.” Granted, I’m sure my Reddit post is a bit spicy, so I need to go back and edit out all the profanity.
TL;DR
1 Not having more partitioning and RAID options is very bad.
2 Yast2 tools not working is very bad.
3 I am officially challenging whatever agama dev use case survey was conducted: who was polled? Not their names or anything PII/DOX, but who as in where were they sourced from? IF (note I am saying “IF”) those polled were only SUSE users, and that was the decision to put agama also on OpenSUSE, that reveals a sampling error and breaks the validity of the research. You can’t say opensuse users will like agama if you polled only suse enterprise users, for example. You can’t say “no one” will install opensuse in whatever scenario if the research was only conducted with suse enterprise users, for example. It’s like the problem that was happening in medicine where prospective pain relievers were only researched on males (and the subsequent way in which women expressed that they don’t always work as well for them etc.).
