Failure to start after upgrade to f41. CUPS and Security authorisation issue
Log says:
[FAILED] Failed to start auditd.service - Security Audit Logging Service.
[FAILED] Failed to start cups.service - CUPS Scheduler.
" " " " " " " " " " " " "
LOOPS OF THE ABOVE REPEAT DOWN THE SCREEN
Any advice much appreciated for any work around on this
Does https://discussion.fedoraproject.org/t/failed-to-start-security-auditing-service/77095/2 match your issue (missing /var/log/audit directory)? You may get different errors – if they don’t lead you to a fix post details here.
If the cups.service still fails to start, you can check systemctl status cups and journalctl -b -u cups to see if they provide helpful details.
Thanks George I will look at the link. The problem now is how to get to terminal prompt to follow suggestions without the computer booting? Can do anything through BIOS options? (I am seeing this on a different device but the Linux fedora computer won’t boot at all - although I can get into BIOS).
You can try booting to a text console by using the grub editor to remove the rhgb quiet from the kernel command line and adding a <space>3 at the end of the command line.
If that isn’t helpful, run journalctl --directory=.... in a Live USB system to view the journals in the system disk. See man journalctl for details. Mount the system “root” directory with Gnome disks and then .... should be <mount_location>/var/log/journal/<long_random_string>/user-<uid>.journal.
How do I boot into a text colsole? Please excuse my ignorance in these matters.
I managed to boot the computer into permissive emergency mode. How do I find out what the file problem is and begin to correct it? I would like the computer to be able to boot normally without having to do a repeat editing of the grub to enter emergency mode each time.
I tried that fix but it did not work - touch was an unrecognised command.
Same for fixfiles command. Neither were recognised.
Then I tried booting within emergency mode again and tried the same commands within terminal - but permission was denied.
There was more success with fixfiles relabel in terminal - command was recognised but again the operation was not permitted on the removal of temp files.
Could not set context for /…: Operation not permitted
systemd-private-…:Permission denied
Is there a way to do this pre-boot within the GRUB or after booting into emergency recovery mode from within Terminal?
That would happen if you tried to run the commands in the grub shell – they are meant to be run in the text console, which is a bash shell in a fresh install of Fedora Workstation.
I have decades of experience introducing new users to Linux, so have seen this confusion before. While many users never need to use a terminal, it is an essential tool for solving problems and has the advantage that it is easy to post error messages as searchable text.
Linux Command is a time-tested reference that has helped many new users with command-line tools. I recommend spending a few hours learning your way around the reference and then referring back to it when you try something new.
I’ve managed to get the computer booted into emergency mode again. I have tried various commands to try to solve the ongoing issues.
I tried a recommended post-upgrade command which resulted in:
Updating and loading repositories:
skype (stable) 100% | 9.0 KiB/s | 1.5 KiB | 00m00s
Status code: 503 for https://repo.skype.com/rpm/stable/repodata/repomd.xml (
Status code: 503 for https://repo.skype.com/rpm/stable/repodata/repomd.xml (
Status code: 503 for https://repo.skype.com/rpm/stable/repodata/repomd.xml (
Status code: 503 for https://repo.skype.com/rpm/stable/repodata/repomd.xml (
Librepo error: Cannot download repomd.xml: Cannot download repodata/repomd.x
Repositories loaded.
Failed to resolve the transaction:
Problem 1: installed package add-determinism-0.3.6-3.fc41.x86_64 requires libc.so.6(GLIBC_2.39)(64bit), but none of the providers can be installed
I previously tried to remove skype which is now redundant, but, for the same reason as above; it resulted in - unable to complete - conflicting requests
Please show us the “post-upgrade” command you used.
christiantaylor@Fedora:~$ sudo dnf install rpmconf
[sudo] password for christiantaylor:
Updating and loading repositories:
skype (stable) 100% | 4.2 KiB/s | 1.5 KiB | 00m00s
Status code: 503 for https://repo.skype.com/rpm/stable/repodata/repomd.xml (
Status code: 503 for https://repo.skype.com/rpm/stable/repodata/repomd.xml (
Status code: 503 for https://repo.skype.com/rpm/stable/repodata/repomd.xml (
Status code: 503 for https://repo.skype.com/rpm/stable/repodata/repomd.xml (
Librepo error: Cannot download repomd.xml: Cannot download repodata/repomd.x
Repositories loaded.
Failed to resolve the transaction:
Package “rpmconf-1.1.10-2.fc39.noarch” is already installed.
Problem 1: installed package kernel-modules-extra-6.5.10-100.fc37.x86_64 requires kernel-uname-r = 6.5.10-100.fc37.x86_64, but none of the providers can be installed
A more successful command was to run: sudo rpmconf -a
This gave me a list of updates to confirm with Y/N 's - I opted Y for all…
christiantaylor@Fedora:~$ sudo mount | grep "/boot "
/dev/sda4 on /boot type ext4 (rw,relatime,seclabel)
[sudo] password for christiantaylor:
/dev/nvme0n1p2 on /boot type ext4 (rw,relatime,seclabel)
bash: syntax error near unexpected token `(’
christiantaylor@Fedora:~$ sudo dnf install clean-rpm-gpg-pubkey
sudo clean-rpm-gpg-pubkey
[sudo] password for christiantaylor:
Updating and loading repositories:
skype (stable) 100% | 6.5 KiB/s | 1.4 KiB | 00m00s
Status code: 503 for https://repo.skype.com/rpm/stable/repodata/repomd.xml
Status code: 503 for https://repo.skype.com/rpm/stable/repodata/repomd.xml
Status code: 503 for https://repo.skype.com/rpm/stable/repodata/repomd.xml
Status code: 503 for https://repo.skype.com/rpm/stable/repodata/repomd.xml
Librepo error: Cannot download repomd.xml: Cannot download repodata/repomd.
Repositories loaded.
Failed to resolve the transaction:
Package “clean-rpm-gpg-pubkey-0-1.20210505gitebb9ab1.fc39.noarch” is already installed.
Problem 1: installed package kernel-modules-extra-6.5.10-100.fc37.x86_64 requires kernel-uname-r = 6.5.10-100.fc37.x86_64, but none of the providers can be installed
rpmconf helps deal with configuration files, but you have packages from Fedora 37 and 39. This can happen when a package is abandoned so there is no upgrade available, but there are current versions of the Fedora 41 kernel modules and rpmconf. It seems you have some old packages built with old libraries that break updating to Fedora 41.
At this point, you may be better off making a backup of user files and local configuration changes (usually edits to configuration files under /etc) and doing a fresh install. If that is not an option (e.g., you depend on a a package that is not available in F41), you would need to make a list of installed packages from older Fedora versions (in a terminal, sudo dnf list installed grep -v fc41). Decide if any are “mission critical”, and try to remove the others. Look for current replacements for “mission critical” packages.
As a last resort, you can create a VM running an older Fedora version, but often old “mission critical” packages are no longer available. If source is available you may be able to build the app.
Thank you George, I think I’ll try a fresh f14 install having first saved my files onto an external disk. It seems by far the easiest option at this stage although I was kind of holding out for some magic terminal command that would succeed in deleting all the older packages leaving only the f41 system behind. Thank you for your patience and kind advice.