PackageKit hangs F40 during update from 6.8.5-301 to 6.8.11-300

[radial-dimension]

Hi there, first time on Linux, first week with Fedora. If there’s any information I can provide let me know.

I’ve been trying to upgrade my kernel from 6.8.5-301 to 6.8.11-300 and failing constantly, always hanging at 97% using Software app:

[    *] Job packagekit-offline-update.service/start running (time / no limit)

or it hangs at Running scriptlet when using sudo dnf upgrade:

  Running scriptlet: kernel-modules-extra-6.8.11-300.fc40.x86_64                                     5/5 
  Running scriptlet: kernel-modules-core-6.8.11-300.fc40.x86_64                                      5/5 
  Running scriptlet: kernel-core-6.8.11-300.fc40.x86_64                                              5/5
  ▓ 

I did some googling and found this post that was nearly identical, but my boot args don’t reference any swap volumes or have resume on them:

BOOT_IMAGE=(hd0,gpt2)/vmlinuz-6.8.5-301.fc40.x86_64 root=UUID=d9f26449-bd1c-4bd3-8bf8-eda386a2c24a ro rootflags=subvol=root rhgb quiet

I am having A LOT of errors with my OCTO device and CoolerControl, but I doubt that’s relevant. No current fix for it either since the are no official drivers as far as I can tell. Here’s an output of sudo journalctl -xe after a failed install and booting up 6.8.5-301 again:

Jun 15 16:51:16 RADI coolercontrol-liqctld[1932]: ERROR    coolercontrol_liqctld.device_service - Error setting fixed speed:
Jun 15 16:51:16 RADI coolercontrol-liqctld[1932]: Traceback (most recent call last):
Jun 15 16:51:16 RADI coolercontrol-liqctld[1932]:   File "/usr/lib/python3.12/site-packages/coolercontrol_liqctld/device_service.py", line 435, in set_fixed_speed
Jun 15 16:51:16 RADI coolercontrol-liqctld[1932]:     status_job.result(
Jun 15 16:51:16 RADI coolercontrol-liqctld[1932]:   File "/usr/lib64/python3.12/concurrent/futures/_base.py", line 456, in result
Jun 15 16:51:16 RADI coolercontrol-liqctld[1932]:     return self.__get_result()
Jun 15 16:51:16 RADI coolercontrol-liqctld[1932]:            ^^^^^^^^^^^^^^^^^^^
Jun 15 16:51:16 RADI coolercontrol-liqctld[1932]:   File "/usr/lib64/python3.12/concurrent/futures/_base.py", line 401, in __get_result
Jun 15 16:51:16 RADI coolercontrol-liqctld[1932]:     raise self._exception
Jun 15 16:51:16 RADI coolercontrol-liqctld[1932]:   File "/usr/lib/python3.12/site-packages/coolercontrol_liqctld/device_executor.py", line 36, in run
Jun 15 16:51:16 RADI coolercontrol-liqctld[1932]:     result = self.fn(**self.kwargs)
Jun 15 16:51:16 RADI coolercontrol-liqctld[1932]:              ^^^^^^^^^^^^^^^^^^^^^^
Jun 15 16:51:16 RADI coolercontrol-liqctld[1932]:   File "/usr/lib/python3.12/site-packages/liquidctl/driver/aquacomputer.py", line 500, in set_fixed_speed
Jun 15 16:51:16 RADI coolercontrol-liqctld[1932]:     self._set_fixed_speed_directly(channel, duty)
Jun 15 16:51:16 RADI coolercontrol-liqctld[1932]:   File "/usr/lib/python3.12/site-packages/liquidctl/driver/aquacomputer.py", line 445, in _set_fixed_speed_directly
Jun 15 16:51:16 RADI coolercontrol-liqctld[1932]:     ctrl_settings = self.device.get_feature_report(_AQC_CTRL_REPORT_ID, report_length)
Jun 15 16:51:16 RADI coolercontrol-liqctld[1932]:                     ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
Jun 15 16:51:16 RADI coolercontrol-liqctld[1932]:   File "/usr/lib/python3.12/site-packages/liquidctl/driver/usb.py", line 495, in get_feature_report
Jun 15 16:51:16 RADI coolercontrol-liqctld[1932]:     data = self.hiddev.get_feature_report(report_id, length)
Jun 15 16:51:16 RADI coolercontrol-liqctld[1932]:            ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
Jun 15 16:51:16 RADI coolercontrol-liqctld[1932]:   File "hidraw.pyx", line 361, in hidraw.device.get_feature_report
Jun 15 16:51:16 RADI coolercontrol-liqctld[1932]: OSError: read error
Jun 15 16:51:16 RADI coolercontrold[2172]: Error applying scheduled speed setting: Setting fixed speed for LIQUIDCTL Device #1: ed0bc39916e0a82efb8013aa3206d16eaa76e91e8fc3ae4d7c65aaae0ef58fde
Jun 15 16:51:16 RADI coolercontrol-liqctld[1932]: WARNING  liquidctl.driver.aquacomputer - required PWM functionality is not available in aquacomputer_d5next kernel driver, falling back to direct access

I’ve tried various dnf cache/package related commands to sort out any dependency issues and reinstalled PackageKit to no avail. sudo dnf clean all sudo dnf makecache sudo rpm --rebuilddb sudo dnf reinstall PackageKit! Any pointers?

[glb (Gregory Lee Bartholomew)]

In this case, I’d try updating the rest of the system first, then see if you can get the kernel update to succeed in a later update run after you’ve finished updating the system and rebooted.

You can exclude the kernel from the system update by adding --exclude=kernel* to your sudo dnf update command.

Also, you should run dnf update (i.e. the “online” version of the update command) from runlevel 3 (the command line, without the full GUI running). One way to get to runlevel 3 is by switching to a VT (e.g. CTRL+ALT+F4), then sign in and run init 3. The last number from the output of the runlevel command will confirm what runlevel you are in (5 is the “graphical” runlevel, 3 is the “multi-user” runlevel).

[radial-dimension]

@glb

–exclude=kernel*

Hi Gregory, the rest of the system has been up to date now after trying a few times, only latest kernels haven’t been updated.

radialdimension@RADI:~$ sudo dnf update --exclude=kernel*
[sudo] password for radialdimension: 
Last metadata expiration check: 1:42:13 ago on Sat 15 Jun 2024 15:54:29 BST.
Dependencies resolved.
Nothing to do.
Complete!

[glb (Gregory Lee Bartholomew)]

My guess is that the kernel-core script is trying to update the contents of your boot partition. Is it possible that your boot partition is full?

Also, do you have anything installed that is trying to run DKMS to dynamically build kernel modules?

[radial-dimension]

I’m guessing 1.0GB (default) is enough for boot partition?

The install is incredibly vanilla right now, so no fancy apps or customisation. Getting fans to work was the main thing. I installed TimeShift after failing the update a few times and needed a fallback option. TimeShift is configured to use an external drive.

I’m running dual boot with Windows but on completely separate drives, secure boot is off.

[glb (Gregory Lee Bartholomew)]

It looks like you have plenty of space in both /boot and /boot/efi.

I haven’t run into this problem before, so I’m not sure what to suggest. I think I’d try to get some debugging output from that kernel-core script. Looking at the man pages, it appears you might be able to do that by passing -vv to the rpm command. So, trying downloading the kernel packages with dnf download kernel*. Delete the “debug” and “uki” kernel packages (ugh, also delete the “kernelshark” package). Then trying installing the remaining kernel packages with rpm -ivv kernel*. Hopefully that double v option will provide some clues about what is going wrong.

Actually, there is a better way to dowload the kernel packages using koji: Installing Kernel from Koji :: Fedora Docs Maybe that will avoid picking up some unnecessary packages. (sudo dnf install on that quick docs page should be sudo dnf install *. I’ve submitted a PR to get that fixed. But in your case, don’t use dnf install ... use the rpm -ivv ... command I provided earlier so you will, hopefully, get some extra feedback about what failed.)

[radial-dimension]

Thanks for taking the time on this. I’m out of my depth but rpm -vvh kernel* and koji worked in the end! It did hang on this line for a while:

+ /bin/kernel-install add 6.8.11-300.fc40.x86_64 /lib/modules/6.8.11-300.fc40.x86_64/vmlinuz

I tried journalctl on another Terminal, no errors but soon after the rest of the install completed.

I set version 6.8.11-300 as default following the Additional Steps: from your first link and rebooting completed the process.

Here’s the terminal copy if it helps.

[hamrheadcorvette (hammerhead corvette)]

@radial-dimension"P.S

Hi there, first time on Linux, first week with Fedora.
I’ve been trying to upgrade my kernel … and failing constantly,

Welcome to , Welcome to

[glb (Gregory Lee Bartholomew)]

@radial-dimension

I’m out of my depth but rpm -vvh kernel*

Sorry about that. I wrote the command somewhat hurriedly and partially from memory as I was going out the door. There should not be an h in the flag list (the h summarizes the installation progress as a series of hash marks progressing across the screen; it is the opposite of what you want if you are trying to see the details of the installation). Also, a lower-case i (for “install”) should be in the flags list. I’ll correct my earlier post.

Glad you got it working (in spite of my instructions ).

@radial-dimension

koji worked in the end! It did hang on this line for a while:

+ /bin/kernel-install add 6.8.11-300.fc40.x86_64 /lib/modules/6.8.11-300.fc40.x86_64/vmlinuz

If it is hanging there for a long time, my guess would be that it is compiling a kernel module. I don’t know what else might cause that. It is normal for kernel installs to take a long time in some configurations. (My PC, for example, normally hangs for a while during kernel installs because I use ZFS which requires a kernel module to be compiled every time the kernel is updated. You might try running dkms status to see if it shows any kernel modules being dynamically built by that system.)