New install of F43 systemctl start cups.service Unknown SystemGroup "sys" on line 22 of /etc/cups/cups-files.conf. Cause: missing sys in /etc/group. Reproducible: Always
Hi, thank you for the report! I got F43 VM and group 'sys' is present by default in setup package What is your Fedora installation which does not include group 'sys'? The values for SystemGroup is generated during build based on groups present on the build machine - I could define them by configure option, but autodetect looks more flexible IMO. The fix is to remove 'sys' from SystemGroup in your /etc/cups/cups-files.conf, but since 'sys' is present in default Fedora Linux, I would keep the default config as it is. I'll close the issue for now - I can reopen if there is Fedora flavor which does not have group 'sys' by default.
FTR, would you mind attaching your /etc/group file?
And is there /etc/group- file on your machine? If there is, please upload it as well. Thank you in advance!
My system was previously running Mageia Cauldron. Are you wanting the /etc/group* from Mageia (always had "sys:x:3:") or from Fedora 43 (no sys after install)?
Fedora 43. Would you mind sharing how you installed that Fedora 43 by steps (what you did/what you set)? Because it is really strange for 'sys' to be missing in default installation.
1. dd'd the / and /boot/EFI partitions with SystemRescue as backup in case. 2. Did a normal install pointing to these 2 partitions; plus /home to 3rd partition. 3. Booted. Most of my stuff just worked; except that I had to install various packages, including cups after which I saw the error in the journal. Then added sys entry into /etc/group which solved the cups failure. Dec 31 15:24:02 systemd[1]: Starting cups.service - CUPS Scheduler... Dec 31 15:24:02 cupsd[2939424]: Unknown SystemGroup "sys" on line 22 of /etc/cups/cups-files.conf. ^^^^^^^^^^^^^^^^^^^^^^^^^ Dec 31 15:24:02 cupsd[2939424]: Unable to read "/etc/cups/cups-files.conf" due to errors. Dec 31 15:24:02 systemd[1]: cups.service: Main process exited, code=exited, status=1/FAILURE Dec 31 15:24:02 systemd[1]: cups.service: Failed with result 'exit-code'. Dec 31 15:24:02 systemd[1]: Failed to start cups.service - CUPS Scheduler. Dec 31 15:24:02 audit[1]: SERVICE_START pid=1 uid=0 auid=4294967295 ses=4294967295 msg='unit=cups comm="systemd" exe="/usr/lib/systemd/systemd" hostname=? addr=? terminal=? res=failed' Dec 31 15:24:03 systemd[1]: cups.service: Scheduled restart job, restart counter is at 1. Dec 31 15:24:03 systemd[1]: Stopped cups.service - CUPS Scheduler. Dec 31 15:24:03 audit[1]: SERVICE_START pid=1 uid=0 auid=4294967295 ses=4294967295 msg='unit=cups comm="systemd" exe="/usr/lib/systemd/systemd" hostname=? addr=? terminal=? res=success' Dec 31 15:24:03 audit[1]: SERVICE_STOP pid=1 uid=0 auid=4294967295 ses=4294967295 msg='unit=cups comm="systemd" exe="/usr/lib/systemd/systemd" hostname=? addr=? terminal=? res=success' Dec 31 15:24:03 systemd[1]: cups.socket: Deactivated successfully. Dec 31 15:24:03 systemd[1]: Closed cups.socket - CUPS Scheduler. Dec 31 15:24:03 systemd[1]: Stopping cups.socket - CUPS Scheduler... Dec 31 15:24:03 systemd[1]: Listening on cups.socket - CUPS Scheduler.
This is the slight difference between the Mageia and Fedora files # diff /mnt/mga/etc/cups/cups-files.conf /etc/cups/cups-files.conf 24a25,27 > # Are Unix domain socket peer credentials used for authorization? > PeerCred on > # ll /mnt/mga/etc/cups/cups-files.conf /etc/cups/cups-files.conf -rw-r-----. 1 root lp 2811 Dec 11 19:00 /etc/cups/cups-files.conf -rw-r----- 1 root lp 2732 Oct 12 07:26 /mnt/mga/etc/cups/cups-files.conf
In case it helps, see https://bugzilla.redhat.com/show_bug.cgi?id=2404966#c42
Ok, IMHO it is not a good practice to copy configuration between different operation systems - it can cause issues which are not bug in software, but configuration error caused by copying, like this one.
If the copying cannot be avoided, it should be done with caution and examine thoroughly any issues whether it is actual bug or configuration issue caused by configuration.
I think you misunderstood something. I did *not* copy from the previous OS; any copying was done by the F43 install process -- I was impressed at how much was gleaned from the old OS saving me much effort in getting the system productive quickly. My point in comment 8 was that the install process and/or installing cups failed to add "sys:x:3:"; yet something else later added that to the end of the root entry, corrupting that file. In my 28 years using Linux exclusively, I have NEVER messed with the _system_'s /etc/group entries (root,sys,...).