During a Livepatch update this morning, something appears to have happened to my laptop.
NVIDIA drivers were not loaded, and Ubuntu loaded a default graphics driver.
Updating to the latest driver (580) made things worse. Now I cannot access kernel 6.8.0-87-generic at all. (see picture)
For the time being, I am on kernel 6.8.0-85-generic.
At the same time, Software Updates wants to force me to do a partial pass, which is bad news.
How do I fix this?
There is one line that worries me a bit:
ALERT! UUID=dfd0efb4-a5ef-4169-a1e2-fa7a8efa62bf does not exist.
Which generally means: it can not find your harddisk, or more specific, your root/boot partition.
The odd thing is that I can boot from Kernel 6.8.0-85, but not 6.8.0-87
If you compare that UUID to the one reported in 6.8.0-85, is it the same as for the failed 6.8..0-87 boot?
Also, have you reviewed all the recent discussions about issues surrounding NVIDIA drivers and possibly freezing or reverting to an older known-good version?
can you open a terminal and execute this command ?
grep UUID /boot/grub/grub.cfg
The long numbers ( following root=UUID= ) should be all the same.
If not, post your output here. 
grep UUID /boot/grub/grub.cfg
linux /boot/vmlinuz-6.8.0-87-generic root=UUID=dfd0efb4-a5ef-4169-a1e2-fa7a8efa62bf ro quiet splash $vt_handoff
linux /boot/vmlinuz-6.8.0-87-generic root=UUID=dfd0efb4-a5ef-4169-a1e2-fa7a8efa62bf ro quiet splash $vt_handoff
linux /boot/vmlinuz-6.8.0-87-generic root=UUID=dfd0efb4-a5ef-4169-a1e2-fa7a8efa62bf ro recovery nomodeset dis_ucode_ldr
linux /boot/vmlinuz-6.8.0-86-generic root=UUID=dfd0efb4-a5ef-4169-a1e2-fa7a8efa62bf ro quiet splash $vt_handoff
linux /boot/vmlinuz-6.8.0-86-generic root=UUID=dfd0efb4-a5ef-4169-a1e2-fa7a8efa62bf ro recovery nomodeset dis_ucode_ldr
linux /boot/vmlinuz-6.8.0-85-generic root=UUID=dfd0efb4-a5ef-4169-a1e2-fa7a8efa62bf ro quiet splash $vt_handoff
linux /boot/vmlinuz-6.8.0-85-generic root=UUID=dfd0efb4-a5ef-4169-a1e2-fa7a8efa62bf ro recovery nomodeset dis_ucode_ldr
Yes, I have read it. I would like to know why all this happened in the first place though. So that I can avoid it in the future. When I booted the laptop this morning, everything was fine. When I booted it later, troubles everywhere. I did not update anything today (intentionally that is). That leaves the Livepatch as main suspect. I have turned it off. Sorry if I sound like I’m rambling, but I’m panicking right now.
From your running -85 OS, can you provide a list of all the files in /boot? I ask this to give you specific guidance on how to make your Old OS the default for the boot menu and "hide" what was installed for the non-working -87.
Also, do you have a full backup from just before the updates occured?
I realize it is late to offer this guidance, but strongly recommend you never perform unattended updates/upgrades. You should perform that only by initiating manually ... after a full backup ... so that nothing you've worked on might be lost by the possible need to recover from full backup. 
Also, by doing manually, you can detect any issues immediately and, if your OS/tools provide the mechanism, revert back "gracefully" after making all the necessary note-taking to document what messages report where things go wrong. Specifically looking at the details/progress/failures documented in each of
This? Oh, and by the way, thanks for finding the time to answer
/boot$ ls
config-6.8.0-85-generic memtest86+.elf
config-6.8.0-86-generic memtest86+_multiboot.bin
config-6.8.0-87-generic System.map-6.8.0-85-generic
efi System.map-6.8.0-86-generic
grub System.map-6.8.0-87-generic
initrd.img vmlinuz
initrd.img-6.8.0-85-generic vmlinuz-6.8.0-85-generic
initrd.img-6.8.0-86-generic vmlinuz-6.8.0-86-generic
initrd.img-6.8.0-87-generic vmlinuz-6.8.0-87-generic
initrd.img.old vmlinuz.old
memtest86+.bin
May want to check the output of:
sudo update-initramfs -c -k 6.8.0-87-generic
If there's a warning or error; that might help us figure out why 6.8.0-87 specifically is failing.
Hey, no error or worning. Output:
update-initramfs: Generating /boot/initrd.img-6.8.0-87-generic
Try sudo update-grub and restart. See if its still failing.
I’ve just tried, it still fails.
I would follow Stephen's guidance until you hit a brick wall.
If that happens, you can "restore" the -85 image as the default by doing
(EDIT: note correction of last line)
mv initrd.img initrd.img.87
mv initrd.img.old initrd.img
mv vmlinuz vmlinuz.87
mv vmlinuz.old vmlinuz
A reboot will show the correct choice of only -85 and its alternates listed, as long as you have the os-prober in your
showing
GRUB_DISABLE_OS_PROBER="false"
but if you run update-grub after that, I am not sure, but I believe it will undo that renaming sequence. Maybe Stephen (@stephematician) can clarify on that specific point.
Okay, thanks for the insight. While trying different things, I noticed this:
apt-mark showmanual | grep nvidia
linux-modules-nvidia-550-generic-hwe-22.04
linux-modules-nvidia-580-generic-hwe-22.04
nvidia-driver-550
nvidia-driver-580
nvidia-prime
Does this mean there are two sets of drivers or just one?
There will be just one set (hopefully).
There's a command to build all the additonal modules. I think it is /usr/lib/dkms/dkms-autoinstaller start but I dont have my UM24.04 machine on hand right now to test.
If something went wrong with building the additional modules for the kernel - that would be good to know.
Otherwise I'd also try something like sudo apt reinstall linux-image-generic (or the 'hwe' image if thats what you use). I would provide better instructions but I'm heading out, good luck for now!
Also re dkms - you need some additional arguments so you only build the modules for 6.8.0-87. Look them up before you run it.
I have already tried to reinstall the 87 kernel, but it did not work.
Just to be clear, what are my best options now? Out of desperation, I was thinking of removing the 86 and 87 kernels and trying to have only the 550 drivers. But I’m afraid it would make an even bigger mess…
If you think its the NVIDIA proprietary driver thats causing issues - remove it (purge even) and use the Nouveau driver.
When you reinstalled - what output did you get from the DKMS build (which should have been triggered with a reinstall)?
My apologies if the question is off-topic. Is there a way to check if some packages were incorrectly/partially installed? I have this feeling that the cause of all these troubles might be connected to that.
Or it’s really because NVIDIA removed the 550 drivers (and only those, see picture) from the list 
