• SupportQuestion

  • Recent update switched to LTS kernel, boot fails, manual steps required.

Not sure if this is a bug or just an issue on my system. When I ran today's update (June 12, 2020) and rebooted, my system hung with the "_" cursor on the screen. I had to force shut it down. Powering up I could see that the EFI boot was selecting the LTS kernel as default instead of the Current one that I've always been running. I manually selected the Current kernel and the system booted fine. I then ran sudo clr-boot-manager update and rebooted. The LTS kernel was still selected as default, and the same issue of booting to the black screen with the cursor. I then removed the kernel using sudo eopkg remove linux-lts linux-lts-headers. I then rebooted, and still the LTS kernel was selected as default. After manually selecting the Current kernel I booted into the system and manually mounted the EFI partition to boot, and after backing it up, I manually removed all the LTS entries from the loader and EFI directories. I then ran sudo clr-boot-manager update again. Finally, when I rebooted the LTS kernel entry was gone and it booted properly into the Current kernel.

I would like to keep the LTS kernel installed for those corner cases where I may need to boot to it, but not if it's going to cause issues with it being selected by default.

Are there any suggestions on how to fix this issue?

Thanks!

    You can use sudo clr-boot-manager list-kernels to get a list of the available kernels and the use sudo clr-boot-manager set-kernel <kernel from list> followed by a sudo clr-boot-manager update to select the -current kernels. Once you make the switch, clr-boot-manager should remember to keep you on -current going forward. Sorry about this. There seems to be something going on with the latest clr-boot-manager that we haven't nailed down yet.

      DataDrake Thanks for the reply.

      I actually tried those steps as well, I just didn’t document them in my description. (I felt I was getting too long winded already! LOL)

      Hopefully we’ll see some fixes for clr-boot-manager soon...

        yep same thing here i updated today

        camtron I had exactly the same problem. Somehow an lts kernel image was still in my boot folder although I didn't have any lts kernel installed. And it seemed like clr-boot-manager always selected that lts kernel over 5.6.18. I don't know how clr-boot-manager works, but I think it somehow "thought" the lts kernel was the newest one?

        In my case it was kernel image 4.9.210-151.lts hanging around in the boot folder.

          clr-boot-manager fix has hit Shannon as of a few minutes ago.

          1. Update for eopkg or the Software Center
          2. Run sudo clr-boot-manager list-kernels to get a list of the available kernels.
          3. Run sudo clr-boot-manager set-kernel <line from kernel list> to set a default kernel (no <> symbols)
          4. Run sudo clr-boot-manager list-kernels to check for the * next to the kernel you selected.
          5. (Optional) Run sudo clr-boot-manager update and sudo clr-boot-manager list-kernels to verify that it did not change after you set it.
            • Note: sudo clr-boot-manager update will always select the highest kernel release number from the same series (e.g. current-5.6.18-155 instead of current-5.6.13-153, but not lts-4.9.227-160).
          6. (Optional) Reboot to confirm that it indeed loaded the entry you selected.

            DataDrake How do I remove the lts kernel from the list? btw, it s/b set-kernel not set kernel right? Thank you for your help! 😀

                DataDrake Note: sudo clr-boot-manager update will always select the highest kernel release number from the same series (e.g. current-5.6.18-155 instead of current-5.6.13-153, but not lts-4.9.227-160).

                Did the update then immediately rebooted. Didn't see the mini-splash featuring the current and older versions, it just simply rebooted and went to the highest kernel. Uname -r confirms this. Success--this probably was exactly as intended.

                  Did the "sudo clr-boot-manager set kernel com.solus.xxxxx.xxxxx.xx"
                  now it sais cannot mount boot device /dev/sda1 on /boot: no such device?

                  elfprince nice catch. Most kernels will be removed automatically on a successful boot for one of the kernels in that branch. If you want to get rid of one entirely, you'll need to mount /boot and remove the initrd, vmlinuz, and loader/entries/<kernel>.conf, then re-run sudo clr-boot-manager update to fix the menus.

                      elfprince It's mostly safe if you're paying attention. Generally, it's only necessary if you have -current and -lts installed and want to remove one of them entirely.

                          DataDrake hmm, mount: /boot: can't find in /etc/fstab. Anyway, is it a must to remove the lts kernel? I have one current and one lts right now. Can they play in peace?

                              elfprince They should be fine together. I was only presenting an option. /boot isn't in /etc/fstab because it gets unmounted after the boot process completes. clr-boot-manager knows how to mount it as needed. You can look in lsblk for the drive though.