Jul 11 08:36:57 localhost systemd[1]: Save Transient machine-id to Disk was skipped because of an unmet condition check (ConditionPathIsMountPoint=/etc/machine-id).
Jul 11 08:36:57 localhost systemd[1]: Rebuild Journal Catalog was skipped because of an unmet condition check (ConditionNeedsUpdate=/var).
Jul 11 08:36:57 localhost systemd[1]: First Boot Complete was skipped because of an unmet condition check (ConditionFirstBoot=yes).
Jul 11 08:36:57 localhost systemd[1]: First Boot Wizard was skipped because of an unmet condition check (ConditionFirstBoot=yes).
This is another attempt to get my machine to work. I am assured, contrary to advice, that my problem is not amdgpu - the guys over there see nothing out of the ordinary.
Because they are the last thing in journalctl before booting hangs to a command line prompt. They suggest that conditions are not being met to continue with the process.
So … because I have those entries, my machine should lock up?
All working here on Leap 15.6.
Jul 14 17:45:08 mach systemd[1]: First Boot Wizard was skipped because of an unmet condition check (ConditionFirstBoot=yes).
Jul 14 17:45:08 mach systemd[1]: First Boot Complete was skipped because of an unmet condition check (ConditionFirstBoot=yes).
Jul 14 17:45:08 mach systemd[1]: Rebuild Journal Catalog was skipped because of an unmet condition check (ConditionNeedsUpdate=/var).
Jul 14 17:45:08 mach systemd[1]: Commit a transient machine-id on disk was skipped because of an unmet condition check (ConditionPathIsMountPoint=/etc/machine-id).
I’d suggest checking logs for “error” text , or “fatal” , or similar
@Cyclonick please describe “hangs to a command line prompt” do you mean a tty login? If so, login as root user and show the output from systemctl get-default
@malcolmlewis
So that is what a tty login is… what I’ve been staring at despairingly for 2 weeks ! I’m learning geek !
Does journalctl call it “virtual console” or “virtconsole” ? I’m trying to get the hang of how the boot process works. First phase would seem to be building the kernel around the available hardware and then that is passed to systemd - I’ve seen it described somewhere as passing into user space.
As far as my journalctl goes tty seems to be active during the kernel stage and handing over to systemd it closes then there are these 4 unmet condition messages, a slew of systemd units / processes exiting, and virtual console restarting… and my tty login.
Unfortunately, the man at gitlab is not an openSuse user, so he couldn’t commit himself, but he thinks it’s a software problem. One thing I did try was to enable sddm to force a boot into Xorg (amdgpu seem to prefer that for diagnosis…) Sddm wasn’t installed - even though the X11 pattern is… Anyway I installed it - no better, really, but the colour of text in the tty was a bit prettier !
My hypothesis is that during the install (I have to use metaphor) there’s a phase when the software “feels its way round” the host machine and decides how it’s going to operate. On this install it encountered the unpatched virtual hub issue. So somehow it needs to be asked to start again…
I got a message that the code from the patch has been put into production - the day after I’d reported back that the patch seemed to work - so if you hadn’t been so quick to put it into package, we might well be still waiting.
WE HAVE LIFT OFF !!
used the instruction and got multi-user.target and a bit of research moved me to try systemctl isolate graphical.target and suddenly I was in !
My install was really minimum, methinks rather too much - no console, and bash doesn’t recognise “sudo”… - so I rebooted (and ended on the tty again) and did set-default graphical.target, rebooted and now I have my login page.
Also, and I don’t know what to do about it, it does not install Xorg properly, you find yourself in an xwayland environment. It says it has installed the X11 pattern, but in fact it’s only a third of the pattern that has been installed. Whether that’s because of my going for minimum I don’t know. Once I’ve sorted out the bash sudo problem I’ll dup (I think there have been two kernel updates since the one I patched) and see if the kernel has been patched and, if it has and the machine starts cleanly, I’ll start again.
@malcolmlewis
I can’t thank you enough for your help (and patience)
I really believe that if you and @Sauerland hadn’t helped with the patch I’d still be looking at an unusable box - I was the only one to report back to say that it worked, so it looks like I was the only one to try it.
