I did an update today - the Kernel was updated to 5.15.12.
After a reboot, I’m unable to login, although I can still login using the previous Kernel 5.15.8.
I’ve taken a photo of the error messages, but don’t know what to do now.
That graphic is hard to read.
If possible, plug in a USB and save the indicated file to that USB. I cannot read the file name, but presumably you can. And I don’t know your computer. But I would guess that a USB formatted to FAT could be used, and just copy the file to there.
And then file a bug report. See:
openSUSE:Submitting bug reports
Could you post the logs from the boot (either from journalctl or the path specified in the bottom)?
From the image I can’t personally make out the upper portion of the screen’s text.
journalctl --no-pager -b -1 -g boot
That will give you console output from the journal for the previous boot (current boot is 0, then counts down), and search for the pattern “boot” in the output.
You can also replace “boot” with other keywords from the error messages.
It looks as if he is stuck in a dracut shell, and might not have the access to use journalctl.
Thanks for the offer of help, but in trying to copy the file to a usb stick, I managed to break it further and end up with no grub menu at all, and unable to boot into anything.
I tried to recover from there, but as I needed a working PC this morning, I decided to do a quick reinstall.
I’m now running with the 5.15.12 kernel with no issues, so a bit baffled at what happened yesterday - possibly a problem with the new Kernel install?
Apologies to all, and thanks for the offers of help.
I actually got the same error but unfortunately in this scenario, my keyboard is not working to do anything to recover. Booting into the previous kernel (5.15.8-1-default) is working for now. Since a re-install seemed to help, I’ll try that later for the kernel-related packages…let’s see whether that helps but it would be good to know whether you @TWolf already tried that and it eventually broke your system for good
No, I never got round to a re-install of the kernel-related packages (tbh not something I thought of doing), I broke grub before I could do that and decided to do a full re-install instead.
Be good to know if it works.
I’m getting the same error, after updating this afternoon on my (old) Lenovo T420. Another more modern machine is fine with the same update.
Someone said that they were able to run with the previous kernel? That sounds great, but I could not see an option for that. The grub “advanced” menu offers to boot in recovery mode, but it does not offer me an earlier kernel. Can anyone tell me how to do this?
Alternatively, I can boot of the installation USB and go to “Recovery”, but although that drops me at a root prompt, I don’t know how to proceed from there.
I can get “journalctl” output in emergency mode, but I’m not sure how to get it “off” the machine. I tried mounting /boot, but it says that “ext4” is unknown because (I guess) no modules are loading(?)
Any suggestions would be appreciated.
The main boot sequence has many messages that say, “BPF: [NNNNNNN] invalid name_offset [NNNNNNN]”.
The “journalctl” output has many similar messages, with “failed to validate module <module name>] BTF: -22”
Looks like a bugzilla opened for this here: https://bugzilla.opensuse.org/show_bug.cgi?id=1194455
Title mentions VM but some of the comments say baremetal impacted also. I hit this on VM myself.
I eventually got it solved (after booting to 5.15.8) by removing 5.15.12 and reinstalling it. Simply selecting a forced update for 5.15.12 alone didn’t help.
Sounds like some postinstall script needs to run, so maybe it is as simple as rebuilding the initramfs but I don’t know.
Older kernels are still available, e.g.: 5.15.8-1.1 & 5.15.12-1.1.
I skipped 5.15.12 on old Core2Duo and AMD GPU, installing Tiwai’s simpledrm 5.15.13 instead. He since replaced it with 5.16.0.
It worked for me to remove the default kernel and reinstall it via zypper
I thought it was only me since the 5300+ 20GB+ update resulted in my root partition becoming full before the installation of the last 3 or so packages, among them selinux-container or something (was removed when doing a dup after logging in with previous kernel, never to be installed again apparently), but the above solution worked for me.