Update to Kernel 6.14.9-300.fc 42.x86_64 42 Workstation Edition will not boot after install

Recently did an update that included the Fedora Linux kernel, the update added a Fedora Linux 6.14.9-300.fc42.x86_64 42 Workstation Edition entry to the boot menu. When making that choice in the boot menu the screen goes to a flashing cursor for several seconds and then begins to show three rectangular shapes in the middle of the screen that usually flash left to right and then it starts loading the operating system. After this update the three rectangular shapes show up for a few seconds, do not flash and the computer screen goes blank yet the computer stays on in nowhere land. The boot menu does still provide an earlier 6.14.6 kernel that still loads the operating system correctly and everything works fine there.

The computer desktop that I use this operating system on is well over ten years old: Dell Optiplex 780 with an Intel Core 2 Duo E8400, 3.0GHZ, upgraded to maximum allowable DDR3 Ram 16GB, also upgraded to NVIDIA GP108 GeForce GT1030 from integrated, and 8 months ago after looking over suggestions on a Fedora web site about improving performance upgraded from a HDD SATA spinning hard drive to a SSD SATA hard drive. Although I have upgraded, the desktop is old, a used refurbished Xmas gift back around 2016. From past experience with older machines that I used Fedora on, eventually they became too old to install any upgrades to and therefore moved to Debian where a latest version that is still supported was installed.

I suspect this desktop may have reached the end of use on Fedora or is there a flaw that can be corrected.? Is it possible to uninstall the latest kernel and reinstall it over again, in case there was some error that occurred during the update, if so how does one go about removing a kernel and then reinstalling? If someone out there knows of some other error that may have caused this situation what is the solution? If Fedora no longer supports a desktop this old I would appreciate someone that is knowledgeable verifying and I will move on to Debian or Mint since 16GB of memory and a 2GB memory add on graphic card should still be useful somewhere. I will keep the older kernel that still runs the operating system yet stop upgrades so that it will not remove the usable one and replace with a newer version kernel that will not work if support for this model machine has expired. This Fedora operating system does run Blender 3.0 very well with the GPUs showing up in CUDA Cycles Render Devices, something that is a rare occurrence on Linux operating systems. Still the need for an operating system that provides security updates is necessary if the desktop is to be used for the internet.

Before you lose hope, consider the fix mentioned in this post:

Fedora 42 dnf update appends ‘rhgb quiet’ to grub command

Hi,
Have you tried uninstalling the nvidia drivers ?.

After removing the nvidia drivers, and removing the nouveau blacklist from grub.cfg, the PC booted all ok using XWayland, and also X11.

Regards,
Shadders.

Hi.
I have an old laptop with intel core 2 duo P8400 with the maximum installable ram: 4gb.
With DE mate it still works fine even with luks2 applied.

Older machines are prone to hardware issues – some readily apparent “showstoppers” and others just causing “random” crashes. It may not be worth investing much effort to fix older systems, but you should
take the time to check for error messages. I always remove the “rhgb quiet” after installing an OS so I see errors while booting, but you can also press <Esc> to see messages. You can also use the grub2 editor to add <space>3 to see if you can boot to a console and use journalctl to get details of the issue.

With lots of “enterprise” grade systems that don’t meet Windows 11 requirements being dumped by large enterprises, this could be a good time for upgrading to a “refurbished” system. You can check a system’s Linux Hardware Database profiles to see what issues affect linux.

Some of your upgrades can be reused. I find that SSD SATA drives work well in USB3 external cases.

I opened the file that follows the path /etc/default/grub and here is what was found within;

GRUB_TIMEOUT=5
GRUB_DISTRIBUTOR=“$(sed ‘s, release .*$,g’ /etc/system-release)”
GRUB_DEFAULT=saved
GRUB_DISABLE_SUBMENU=true
GRUB_TERMINAL_OUTPUT=“console”
GRUB_CMDLINE_LINUX=“rhgb quiet rd.driver.blacklist=nouveau,
nova_core modprobe.blacklist=nouveau,nova_core”
GRUB_DISABLE_RECOVERY=“true”
GRUB_ENABLE_BLSCFG=true

I do see “rhgb quiet” in this coding yet do not understand if this is an execution. If it is an execute then I need to add through a terminal the line “sudo grub2-mkconfig > /boot/grub2/grub.cfg”?
I apologize for being confused yet this stuff is far above my pay grade, I am an older retired +70 years type with a part-time job as a cook in a restaurant to keep myself above the poverty income level and have a little spending money for this hobby.

One thing I noticed in this file is the “GRUB_TIMEOUT=5” as the first line. I normally change the 2 timeout settings in the grub.cfg to 60 or more so that I have a longer time period to view the boot menu where I normally have several operating systems listed. Changing the grub.cfg to 60 for timeout did not effect the first line in this /etc/default/grub file.

Another question is what does “rd.driver.blacklist=nouveau” mean?
I do recall nouveau being a generic driver for Nvidia graphic cards, I have tried manually installing proprietary drivers from AMD and Nvidia in the past on Debian while at the same time never needed to install them on Fedora years ago.

Hi,
The PC i had the issue with had the nvidia driver installed, but the kernel installation process failed due to konsole crashing during the update. I then had the issue of a blank screen after the three blocks.

From what i understand, you get the three blocks as the system thinks it is using the nvidia driver. I could be wrong on this.

If it cannot implement the nvidia driver then it will default to the nouveau driver, but since you have :

rd.driver.blacklist=nouveau, nova_core modprobe.blacklist=nouveau,nova_core

In your grub.cfg, it cannot implement the nouveau driver.

To check this, at boot time when you see the kernel to boot selected, press the “e” key and you will be able to see the boot grub config. If the above is in the text, select the relevant line and delete the above text only but not the nova_core, then press the F10 key to boot.

Maybe others can state what nova_core does and if it is to be deleted too.

From here you should not just get the three blocks then a blank screen.

If this works, then you know your system has at some stage, implemented the nvidia driver, but an update or upgrade has failed in some way.

Form here, you need to modify the relevant grub config file and use grub2-mkconfig command to update.

I have two fedora system, and each are different - hence you may need to ask what to do to determine the correct process for you.

Regards,
Shadders.

I think it is still under development: kernel.org nova NVIDIA GPU, Nova GPU Driver, but we don’t know if this came from some 3rd party package, so more than deleting a kernel command-line option may be needed.

Maybe modinfo nova_core will help decide what needs to be done.

nova_core is part of the 6.15.x kernel. Recent rpmfusion nvidia driver packages have added nova_core to the blacklist. Don’t remove it.

I have Fedora 42 Workstation Edition on a second machine and right now am writing this with the same kernel installed and running fine 6.14.9. This machine in the boot directory has three kernels that also appear in the boot menu: 6.14.8, 6.14.9 and 6.14.11. For some reason that I do not understand this desktop is keeping three choices of booting kernels instead of only 2 in the machine I am having trouble with 6.14.6 (which still works) and 6.14.9 that is not accessible and has the problem. This machine that is running fine with all three kernels is not a brand name machine, a custom build out of used refurbished parts, an Intel I7 sixth generation (just missed the requirements to upgrade from Windows 10 to 11) yet I am ready with Fedora and Mint installed. Two primary differences with these machines is this one was upgraded with an AMD Radeon 6600 compared to the other machine that has a Nvidia graphic card. Another major difference is the Dell I am having trouble with has a motherboard architecture designed for a Legacy CMOS bios that is Phoenix A14 version last version update back in 2012 and not EFI capable, whereas this custom build has UEFI.

I looked in all three kernels of this newer custom build desktop and the etc/default/grub file reads different than on the older Dell desktop:

GRUB_TIMEOUT=5
GRUB_DISTRIBUTOR=“$(sed ‘s, release .*$,g’ /etc/system-release)”
GRUB_DEFAULT=saved
GRUB_DISABLE_SUBMENU=true
GRUB_TERMINAL_OUTPUT=“console”
GRUB_CMDLINE_LINUX=“rhgb quiet”
GRUB_DISABLE_RECOVERY=“true”
GRUB_ENABLE_BLSCFG=true

On the “GRUB_CMDLINE_LINUX=” line there is a difference compared to the other desktop that includes on this line “rd.driver.blacklist=nouveau, nova_core modprobe.blacklist=nouveau,nova_core”. Is this additional line due to the fact that the Dell desktop has a Nvidia graphic card whereas this one has a Radeon AMD? If I remove that part of the line what will happen, do I risk the possibility of losing the one kernel that is still loadable and opening up the operating system, another possibility being losing the GUI?
Maybe I could save a copy of the two files I test for changing grub.cfg and /etc/default/grub as they are now and if something goes drastically wrong use a live usb version of Fedora Workstation 42 that I have on a thumb drive to put the saved files back in place as they were before trying this? Before trying this I should change the permissions in the path that leads to these two files since I am not sure that from a live version I would be able to do anything other than read the files unless I change them to chmod 777? Then again I have had past experiences playing around with this stuff where I could not do anything to get into certain files or directories and could not bypass permission denied, even with root privileges and password.

I went into the Boot Directory of this 6.14.6 kernel operating system /boot/grub2/ and looked through the grub.cfg file, found that line of code in this area;

if [ -z “${kernelopts}” ]; then
set kernelopts="root=UUID=ab1xy74a-40d5-79z48-1234-9831466l735f ro
#(I have changed the actual UUID after thinking it may be a security risk being exposed to anyone that views this post, no offense to anyone being helpful here, simply do not know if a potential risk is there).
rootflags=subvol=root rhgb quiet rd.driver.blacklist=nouveau modprobe.blacklist=nouveau "
fi

Would this be the correct area to remove the line entries you suggest?

man dnf.conf and read description of installonly_limit

first check whether /boot is large enough to store additional kernels.

Spent 06/23/25 trying a lot of suggestions here. After removing the rd.driver.blacklist=nouveau noave_core modprobe.blacklist-nouveau.nove_core from /etc/default/grub and also the same line entry that I found in /boot/grub2/grub.cfg I restarted the desktop and when attempting to boot the 6.14.9 kernel got a blank screen yet was able to switch to the second, third, or fourth virtual console and use text commands.
Tried a few other things yet nothing worked to get back a GUI yet removing the coding at least allowed me to have the use of command line text, this did not change anything in the earler kernel where I was able to use GUI 6.14.6. Today I started up the 6.14.9 kernel and tried some updates, noticed a new kernel was available along with quite a lot of other updates that were not there a few days ago after updating. After updating I restarted the desktop and a new kernel appeared in the menu, 6.15.3 along with the 6.14.9 that was moved down to the second entry. Tried both kernels yet was not able to get either started, no virtual console was available to switch to Ctrl ALT F2 or anything else just a dead empty screen. Tried a second time an hour or so later and this time although very slow to start both kernels booted me into the Fedora Workstation operating systems with a GUI.

So my several hours of effort trying some suggestions here did at least get me the ability to use a text editor or command line terminal that a couple of days later allowed me to get updates that solved the problem through the terminal command line. Thank you to everyone here for your time and consideration helping out with troubleshooting.