Package update (kernel included) frozen at "Running post-transaction hooks: (1/9) Reloading system manager configuration..."

Hi everyone,

About 15 mins ago I ran yay and noticed several updates including but not limited to kernel & LTS kernel updates. I made a Timeshift backup and then re-ran yay.

It usually goes well, but this time my entire computer is frozen at the “Running post-transaction hooks: (1/9) Reloading system manager configuration…” phase. Pointer is nowhere to be found, gestures are not working, but I can hear the fan spinning dynamically so I think it is not 100% frozen.

Anyway, here are the packages updated in that frozen batch:

gcc-libs, libelf, linux, acpi_call, binutils, mpfr, gcc, lib32-glibc, lib32-glibc-libs, libsysprof-capture, libtool, linux-headers, linux-lts, linux-lts-headers.

(Sorry for improper formatting, I am typing on mobile :sweat_smile:)

The only change I have recently made to the computer is replacing tlp with power-profiles-daemon, about 3 hours before running yay. I have also rebooted my machine before the update.

I had been running from kernel version 6.1.6-arch1-1, and updates to arch1-3. Edit: I forgot to mention I am using systemd-boot.

Now it has been 20 minutes and it is still stuck. Fan is also still spinning. I have pugged it to power so it does not run out of battery for the time being. For reference, kernel updates process usually take less than 10 minutes. Can anyone help advise what I should do? Is it safe to force shut down using the power button? (Device is Lenovo T480s.)

If worse comes to worst I can use the Timeshift backup, but I would like to know what is the “correct” course of action here. Please let me know if you need further information.

Thank you!

Mel

I would chroot into your system first and run the update before using timeshift.

When uninstalling TLP, you have to exactly reverse these steps:
https://linrunner.de/tlp/installation/arch.html

Uncomplete reversal of the TLP installation just might be causing your issue, me guesses. Have a special look at unmasking of these services:
systemd-rfkill.service systemd-rfkill.socket

It may have happened to many, so you’re not alone. :wink:

Okay, I think I am in. I was able to mount my root partition (/dev/nvme0n1p3) as /mnt, my boot partition (/dev/nvme0n1p1) as /mnt/boot, and run arch-chroot /mnt from the live media. It’s no longer showing [root@EndeavourOS liveuser]# but rather [root@EndeavourOS /]#.

Do I just run yay from here? Can you advise me what to do next? I am an uh, meganoob. :sweat_smile:

Follow-up question as an edit: The EndeavourOS guide on chroot-ing used grub instead of systemd-boot. Would I have to mount the boot partition differently?

Yes, I made sure to do them all! :+1:

Yes just run yay

Thank you! Follow up question (edited too late above): the EOS chroot guide I was referring to uses grub by default. As I use systemd-boot, do I have to mount differently?

No you should be fine if you are using EXT4

I was able to run yay after removing the pacman lock file that was preventing it to run. I rebooted and now the kernel entries are back in the systemd-boot menu, however picking any returns a message in red saying (kernel) unsupported or something along that lines… uhh :sweat_smile:

Do I run timeshift from here?

If it’s possible to TTY into your system I would try reinstalling the kernel, if not then this is as far as my Linux wisdom goes my fellow noob.

As of now I am still able to return to the live media, but any attempts to reinstall kernel returns multiple module lines existing in filesystem and “Errors occurred, no packages were upgraded”. :face_with_spiral_eyes:

It is 5 AM where I live and I think I will take a break in 20 mins. If by noon I cannot get this to work, I will most likely reinstall and restore from Timeshift or start afresh with my existing /home. :sweat_smile:

mkinitcpio -P might do the trick?

I would like to try, but I just ran pacman -Q | grep linux and no mentions of linux or -headers kernels are found, also for the LTS. Part of the error messages I get when attempting to pacman -S linux linux-headers (as well as using --overwrite) are as attached:

[...]
linux-headers: /usr/lib/modules/6.1.6-arch1-3/build/sound/soc/sof/mediatek/Kconfig exists in filesystem
linux-headers: /usr/lib/modules/6.1.6-arch1-3/build/sound/soc/sof/xtensa/Kconfig exists in filesystem
linux-headers: /usr/lib/modules/6.1.6-arch1-3/build/sound/soc/spear/Kconfig exists in filesystem
linux-headers: /usr/lib/modules/6.1.6-arch1-3/build/sound/soc/sprd/Kconfig exists in filesystem
linux-headers: /usr/lib/modules/6.1.6-arch1-3/build/sound/soc/sti/Kconfig exists in filesystem
linux-headers: /usr/lib/modules/6.1.6-arch1-3/build/sound/soc/stm/Kconfig exists in filesystem
linux-headers: /usr/lib/modules/6.1.6-arch1-3/build/sound/soc/sunxi/Kconfig exists in filesystem
linux-headers: /usr/lib/modules/6.1.6-arch1-3/build/sound/soc/tegra/Kconfig exists in filesystem
linux-headers: /usr/lib/modules/6.1.6-arch1-3/build/sound/soc/ti/Kconfig exists in filesystem
linux-headers: /usr/lib/modules/6.1.6-arch1-3/build/sound/soc/uniphier/Kconfig exists in filesystem
linux-headers: /usr/lib/modules/6.1.6-arch1-3/build/sound/soc/ux500/Kconfig exists in filesystem
linux-headers: /usr/lib/modules/6.1.6-arch1-3/build/sound/soc/xilinx/Kconfig exists in filesystem
linux-headers: /usr/lib/modules/6.1.6-arch1-3/build/sound/soc/xtensa/Kconfig exists in filesystem
linux-headers: /usr/lib/modules/6.1.6-arch1-3/build/sound/sparc/Kconfig exists in filesystem
linux-headers: /usr/lib/modules/6.1.6-arch1-3/build/sound/spi/Kconfig exists in filesystem
linux-headers: /usr/lib/modules/6.1.6-arch1-3/build/sound/synth/Kconfig exists in filesystem
linux-headers: /usr/lib/modules/6.1.6-arch1-3/build/sound/usb/Kconfig exists in filesystem
linux-headers: /usr/lib/modules/6.1.6-arch1-3/build/sound/usb/line6/Kconfig exists in filesystem
linux-headers: /usr/lib/modules/6.1.6-arch1-3/build/sound/virtio/Kconfig exists in filesystem
linux-headers: /usr/lib/modules/6.1.6-arch1-3/build/sound/x86/Kconfig exists in filesystem
linux-headers: /usr/lib/modules/6.1.6-arch1-3/build/sound/xen/Kconfig exists in filesystem
linux-headers: /usr/lib/modules/6.1.6-arch1-3/build/tools/bpf/resolve_btfids/resolve_btfids exists in filesystem
linux-headers: /usr/lib/modules/6.1.6-arch1-3/build/tools/objtool/objtool exists in filesystem
linux-headers: /usr/lib/modules/6.1.6-arch1-3/build/usr/Kconfig exists in filesystem
linux-headers: /usr/lib/modules/6.1.6-arch1-3/build/version exists in filesystem
linux-headers: /usr/lib/modules/6.1.6-arch1-3/build/virt/kvm/Kconfig exists in filesystem
linux-headers: /usr/lib/modules/6.1.6-arch1-3/build/virt/lib/Kconfig exists in filesystem
linux-headers: /usr/lib/modules/6.1.6-arch1-3/build/vmlinux exists in filesystem
linux-headers: /usr/src/linux exists in filesystem
Errors occurred, no packages were upgraded.

Preceding errors are similar, mentioning /usr/lib/modules/.... Would it make sense for me to rename it into something like, /usr/lib/modules_old/... or completely move it to an extra partition so the kernel installs come through?

Might be worth another try. If a reinstall is looming.

@dalto?

I was able to do a successful LTS kernel install after renaming the kernel in /usr/lib/modules/[lts kernel] into deprecated-[lts kernel] as well as /usr/src/linux-lts into deprecated-linux-lts.

The dracut hook also seemed to run successfully. But upon reboot and trying to boot into the newly installed LTS kernel, I get a kernel panic that starts with Failed to execute /init (error -2) and ends with [ end Kernel panic - not syncing: No working init found. Try passing init= option to kernel.

I… must have made some progress? But I also do not know what to do. Send help. :sos: :sweat_smile:

Edit: I also tried to run mkinitcpio -P as you have advised, but bash cannot find the mkinitcpio command nor can I find it in pacman -Q. :thinking:

You should delete or relocate the files out of /usr/lib/modules. Having an extra kernel and modules in there will produce some strange results.

I would do that and then run sudo reinstall-kernels if you are using systemd-boot.

You are using dracut so the command is dracut-rebuild

Hmm, I still have the same error, both trying to boot into the 6.1.6 kernel and the LTS kernel. :sweat_smile:

I have removed the unnecessary kernels and did a “successful” reinstall of 6.1.6 kernel. The dracut hooks were working post-install, too.

Do I have the wrong boot partition mounted? I used mount [boot partition path] /mnt/boot:face_with_spiral_eyes:

Ehh…did you mount your ESP at /boot? If you are using systemd-boot, that should be mounted at /efi.

You probably need to chroot back in with everything mounted properly and run reinstall-kernels.

On it! Just to be clear, I would simply run mount /dev/... /mnt/efi instead of /mnt/efi, right?

If you are outside the chroot, then yes, you would mount it at /mnt/efi after mounting your root on /mnt

Yes, I did just that :+1:t3:

No luck still, I am getting the exact same kernel panic even after running reinstall-kernels. I will try to rinse and repeat from deleting kernels.