Stream and NVidia driver

Issues related to hardware problems
Post Reply
hhue13
Posts: 1
Joined: 2022/01/20 05:57:42

Stream and NVidia driver

Post by hhue13 » 2022/01/20 06:43:58

I'm having a similar issue on my CentOS8 instlalation with the NVIDIA driver. This worked well until kernel 4.18.0-358 was installed. When starting with kernel 4.18.0-358 the Xorg.0.log shows:

Code: Select all

[    66.761] (II) Loading sub module "ramdac"
[    66.761] (II) LoadModule: "ramdac"
[    66.761] (II) Module "ramdac" already built-in
[    66.776] (EE) NVIDIA: Failed to initialize the NVIDIA kernel module. Please see the
[    66.776] (EE) NVIDIA:     system's kernel log for additional error messages and
[    66.776] (EE) NVIDIA:     consult the NVIDIA README for details.
[    66.776] (EE) No devices detected.
[    66.776] (EE) 
Fatal server error:
[    66.776] (EE) no screens found(EE) 
[    66.776] (EE) 
Please consult the The X.Org Foundation support 
         at http://wiki.x.org
 for help. 
[    66.776] (EE) Please also check the log file at "/var/log/Xorg.0.log" for additional information.
[    66.776] (EE) 
[    66.780] (EE) Server terminated with error (1). Closing log file.
Booting the older `4.18.0-348.7.1.el8_5.x86_64` kernel makes the NVIDIA graphics card again. Any idea what goes wrong here?

Thanks

The link to the full Xorg.0.log: https://www.dropbox.com/s/1vddaohkcb3vi ... 0.log?dl=0
The link to the getinfo.sh output: https://www.dropbox.com/s/ikedwfxifm8yp ... Yo0LF?dl=0

User avatar
TrevorH
Site Admin
Posts: 33191
Joined: 2009/09/24 10:40:56
Location: Brighton, UK

Re: Stream and NVidia driver

Post by TrevorH » 2022/01/20 10:49:55

Check for errors in /var/log/messages and/or /var/log/secure
The future appears to be RHEL or Debian. I think I'm going Debian.
Info for USB installs on http://wiki.centos.org/HowTos/InstallFromUSBkey
CentOS 5 and 6 are deadest, do not use them.
Use the FAQ Luke

User avatar
jlehtone
Posts: 4523
Joined: 2007/12/11 08:17:33
Location: Finland

Re: Stream and NVidia driver

Post by jlehtone » 2022/01/20 11:02:18

How did you install the NVidia driver?

The kernel module (nvidia.ko) has to be (re)compiled for each kernel version.

gjakob
Posts: 16
Joined: 2019/12/22 09:14:23

Re: Stream and NVidia driver

Post by gjakob » 2022/01/21 06:41:21

hhue13 wrote:
2022/01/20 06:43:58
I'm having a similar issue on my CentOS8 instlalation with the NVIDIA driver. This worked well until kernel 4.18.0-358 was installed. When starting with kernel 4.18.0-358 the Xorg.0.log shows:

Code: Select all

[    66.761] (II) Loading sub module "ramdac"
[    66.761] (II) LoadModule: "ramdac"
[    66.761] (II) Module "ramdac" already built-in
[    66.776] (EE) NVIDIA: Failed to initialize the NVIDIA kernel module. Please see the
[    66.776] (EE) NVIDIA:     system's kernel log for additional error messages and
[    66.776] (EE) NVIDIA:     consult the NVIDIA README for details.
[    66.776] (EE) No devices detected.
[    66.776] (EE) 
Fatal server error:
[    66.776] (EE) no screens found(EE) 
[    66.776] (EE) 
Please consult the The X.Org Foundation support 
         at http://wiki.x.org
 for help. 
[    66.776] (EE) Please also check the log file at "/var/log/Xorg.0.log" for additional information.
[    66.776] (EE) 
[    66.780] (EE) Server terminated with error (1). Closing log file.
Booting the older `4.18.0-348.7.1.el8_5.x86_64` kernel makes the NVIDIA graphics card again. Any idea what goes wrong here?

Thanks

The link to the full Xorg.0.log: https://www.dropbox.com/s/1vddaohkcb3vi ... 0.log?dl=0
The link to the getinfo.sh output: https://www.dropbox.com/s/ikedwfxifm8yp ... Yo0LF?dl=0
It looks like that you do not have the nvidia kernel modules installed for that specific new kernel version.
In order to avoid unfortunately regularly occurring situations that a new kernel version gets released in the CentOS Stream repo, but at the same time the repo, from which you get the nvidia drivers, does not provide any precompiled kmod package for that kernel version yet, I recommend to go for any dkms module stream from the nvidia developer repository.

To install the nvidia driver from this repo, you must first remove your already installed nvidia driver. Also make sure that you have the EPEL repository enabled on your system. If not, enable it. Afterwards, install the nvidia developer repo in the following way:

Code: Select all

cd /etc/yum.repos.d
sudo wget http://developer.download.nvidia.com/compute/cuda/repos/rhel8/x86_64/cuda-rhel8.repo
The following command should give you a list of the various module streams provided by that repo.

Code: Select all

dnf module list nvidia-driver
Select any of the dkms streams. The default stream is "latest-dkms". If you want to go for that just type:

Code: Select all

sudo dnf module install nvidia-driver
This command should automatically install all required dependencies (dkms,the "devel" and "header" packages for your running kernel version) in order to enable dkms to automatically build and install the nvidia driver module for your specific kernel.

Post Reply