Exit X before installing

I need to know how to exit the X server in order to download NVIDIA driver that I need to run PixInsight. Currently PixInsight won’t run with the the nouveau graphics drivers. When I try to install the driver, the installation stopped and warned me that I need to exit the X server. I’ve read NVIDIA Read Me and it gave instruction to write a new file to turn it off, but no instruction on how to turn it back on. I rather consult with you before I do anything.

I’m using Fedora KDE Plasma.

Exiting the xserver should not be not required.
Install the nvidia drivers from rpmfusion as shown here Configuration - RPM Fusion and here Howto/NVIDIA - RPM Fusion .
Once that is done then wait about 5 minutes before rebooting.
Verify that secure boot is disabled with mokutil --sb-state. If it is disabled then a reboot should properly load the drivers. If not already disabled then during the reboot you can disable it by entering the bios setup and turning it off. Secure boot must be disabled for the default nvidia drivers to load properly. (One may also sign the modules themselves to enable loading with secure boot enabled)

1 Like

Thanks for your suggestion Jeff,

From PI forum, it was suggested to use this driver.

I just read a little from RFP Fusion, little steep for me at the moment. Will it work with the driver in the link?

The RPM Fusion people package the nvidia code so that it works on fedora.
Use their version and test your app, it will very likely work. As you update
your system you will get updated nvidia drivers automatically.

Note that if you installed the drivers from nvidia then you should remove the same before installing the drivers from rpmfusion. There is a potential for lingering files that are installed differently to cause conflict.

Most installation packages also include the ability to remove the software so try that before installing from rpmfusion. I have not used an nvidia ‘.run’ file for some time so am not familiar with how to uninstall it.

I don’t believe the Nvidia .run driver was installed, since it stopped installing because X server was running.

Now for installing RFM Fusion, do I need to install both free and nonfree repositories?

sudo dnf install https://mirrors.rpmfusion.org/free/fedora/rpmfusion-free-release-$(rpm -E %fedora).noarch.rpm https://mirrors.rpmfusion.org/nonfree/fedora/rpmfusion-nonfree-release-$(rpm -E %fedora).noarch.rpm

After that, I assume to reboot and from what Barry posted, the drivers would be loaded on reboot.

To exit X server

sudo systemctl isolate  multi-user.target

To undo reboot or run

sudo systemctl isolate  graphical.target

But it is better to install nvidia driver from rpmfusion .

1 Like

Thanks for your advise Barry. In the post that follows, I assumed the from home directory, I can install RPM Fusion with the following command lines;

sudo dnf install https://mirrors.rpmfusion.org/free/fedora/rpmfusion-free-release-$(rpm -E %fedora).noarch.rpm https://mirrors.rpmfusion.org/nonfree/fedora/rpmfusion-nonfree-release-$(rpm -E %fedora).noarch.rpm

Then I reboot and the Nvidia drivers will be installed. I believe the command for reboot is # reboot.

As you can see, I’m new at this.

First you setup the rpmfusion repos that you need.

Follow this Configuration - RPM Fusion

Then you can install the nvidia drivers for your card.

Follow Howto/NVIDIA - RPM Fusion

I loaded the free and non-free repositories and loaded the drivers.

The latest driver is 530.41.03 as per the command modinfo -F version nvidia.

Now once I logged into Fedora, the wireless mouse is now erratic, hard to control and PixInsight won’t load.

On the other part of my dual boot, Windows 11, wireless mouse works fine.

We need to know for certain the nvidia drivers are loaded. lsmod | grep nvidia will answer that.
We also need to know what the kernel command line is.
cat /proc/cmdline will provide that.

Once this info is available more pointers may be available.
In particular, the command line probably should include the nvidia-drm.modeset=1 option and the nvidia drivers must be loaded.

[fpaulic@fedora ~]$ su -
Password:
[root@fedora ~]# lsmod | grep nvidia
nvidia_drm 94208 61
nvidia_modeset 1290240 33 nvidia_drm
nvidia_uvm 3219456 0
nvidia 55853056 2991 nvidia_uvm,nvidia_modeset
video 77824 2 asus_wmi,nvidia_modeset
i2c_nvidia_gpu 16384 0
[root@fedora ~]# cat /proc/cmdline
BOOT_IMAGE=(hd2,gpt5)/vmlinuz-6.2.15-300.fc38.x86_64 root=UUID=3a286faa-9d27-4ba4-b941-596284bd98f9 ro rootflags=subvol=root rhgb quiet rd.driver.blacklist=nouveau modprobe.blacklist=nouveau
[root@fedora ~]#

Ok, neither of those commands required that one be root to run them, but not a problem.

If you want to test the effect of adding the nvidia-drm.modeset=1 into the command line for one boot only then it is easy to do by editing the grub command line during boot.
When the grub menu displays press e to edit and on the line that begins with linux add that option into the options shown. Then ctrl+X will continue booting and you can test to see if there are any changes to the items you noted as not working properly.

This change is for the one boot only and is temporary until it is actually added to the permanent grub config, but allows testing of options to see the results.

It didn’t change any thing, I still have problems.

grub


Actually you added a new line with that content.
I suggested that it needed added to the line that begins with linux

Got errors adding nvidia-drm.modeset=1


Are you running as X11 or wayland? I cannot use wayland with nvidia and see the mouse move slowly under wayland, as well as kernel panics.
If so tey useing X11, tou can pick that feom the cogwhell on the lower left(?) on the lovin screen.

How do I now which one I’m running? Are you saying to select X11 from settings?

I remember something this morning. I didn’t do a dnf update before installing drivers.

That time you added it to the initrd line.
Instead it should be added to the linux line.