5.16 Kernel Stuck At Boot

First boot with the new 5.16 kernel has been stuck on the first boot line for 9 hours. Older kernel claims to not be able to load efi. The download and update was all done through the gui. I made sure to -Syu in the terminal for regular packages though.

What can be done to either “fully update” this new kernel, or restore my old one (5.13)?

How to boot from USB into live Manjaro (for repair):

How to increase your chances of solving your issue:

Please provide Information:

Actually that kernel is EOL, so you will not fix it.
Kernel 5.15 would be my recommendation. Not sure why people jump directly to the newest kernel either while they keep an EOL one …
You can install 5.15 from TTY via booting in runlevel 3 with your curent kernel, or from Live media and chroot the installed system.
Let us know if you need any pointers on how to do that.

When I installed 5.16, I was under the impression that the “rc” version was 5.17 now, and that 5.16 is “stable”. The issue I see is that the cursor is stuck blinking after that first line booting in 5.16. However, I am able to interact with the terminal of 5.13. Would I be able to execute my commands from there? If so, what would you suggest?
Because this is my main machine, burning an iso would requure using a friends computer. I would prefer to do what I can from my computer. But if a recovery disk is required, I can manage to make one.
Thanks in advance

Ah, ok. Is understandable.

If you have acces from TTY, is perfect, i actually prefer it over the chroot, so then run this commands as your user:

sudo pacman-mirrors -f5 && pamac upgrade

sudo mhwd-kernel -i linux515

The you reboot by systemctl reboot and while in GRUB menu, go to Advanced and select booting the 5.15 kernel. If that will still fail, then go again to TTY, but this time on the 5.15 kernel and do:

sudo mhwd-kernel -r linux513
and then
sudo mhwd -f -i pci video-nvidia

so with that we are sure the proper kernel modules and driver for nvidia are installed properly.

Just tried it out, seems like each command is failing somewhat due to no name resolution. I would assume having the 5.13 terminal in “Emergency Mode” likely wouldn’t provide the best experience. I think using chroot might be a better option in this case, it is just going to take me about an hour to get it burned. Unless I can open a terminal in the 5.16 kernel, which I highly doubt.

I have occasionally run into this problem when changing kernels. For me simply powering off and on fixes it - all subsequent reboots and fsck work as expected.

@rubiksolver007 you may be able to get to a TTY by pressing CTRL+ALT+F2 on the 5.16 kernel

sounds to me like networking isn’t getting configured at boot, are you on WiFi or can you connect via Ethernet?
Do you know the correct Default Gateway & Subnet settings for your network?

Thanks Bogdan. I was able to install 5.15, but on the last step in tty I get a “breaks dependancy” for nvidia-utils

you need to uninstall cuda

pamac remove cuda

you can reinstall it afterwards :slight_smile:

Thank you very much, both Bodgan and Person. Boot worked with 5.16 using the commands in tty, and making sure to remove cuda. I will mark this as solved.

@rubiksolver007 feel free to reinstall cuda, it’s an optional dependency for blender to speed up render times on your GPU (at least that’s why it’s installed on my system)
it was probably there for a good reason, just the driver can’t be reinstalled with it present

Already did, it was just crucial to remove it in tty. Thanks

This topic was automatically closed 2 days after the last reply. New replies are no longer allowed.