Monitor not detect

[almondon]

Hi All,

I'm using centos 7, I had problem with my monitor and it was not detecting and cannot chage the resolution.
I tried several ways, among them, 1 was good, but the problem was that it was creating seperated kernel
(/usr/lib/modules/6.1.7-1.el7.elrepo.x86_64/kernel),
and for some other problems I need to restart back to other kernel
(/src/kernels/3.10.0-1160.81.1.el7.x86_64)
that still does not recognize the monitor:

https://www.linuxquestions.org/question ... 175583712/
sudo rpm --import https://www.elrepo.org/RPM-GPG-KEY-elrepo.org
sudo rpm -Uvh http://www.elrepo.org/elrepo-release-7. ... noarch.rpm
sudo yum install yum-plugin-fastestmirror
sudo yum --enablerepo=elrepo-kernel install kernel-ml

Is there any way to solve the monitor issue in the same kernel (3.10.0-1160.81.1.el7.x86_64)

Cheers

[TrevorH]

Post the output from lspci -nn | grep -i vga

[almondon]

Thanks for your reply and sorry for being so late.
I have not received email (or to different mailbox.)
I noticed today, the results is as:
01:00.0 VGA compatible controller [0300]: NVIDIA Corporation Device [10de:220a] (rev a1)

Also the size from elrepo is:
"3840x2160_60.00", connected through DP-3.

and I receive following errors, when I try add/new:
$xrandr --newmode "3840x2160_60.00" 712.75 3840 4160 4576 5312 2160 2163 2168 2237 -hsync +vsync
xrandr: Failed to get size of gamma for output default
$ xrandr --addmode DP-3 "3840x2160_60.00"
xrandr: Failed to get size of gamma for output default
xrandr: cannot find output "DP-3"

Cheers

[jlehtone]

The 10de:220a is a RTX 3080 12GB? Do you have NVidia's driver for it (for the normal kernel)?

The open source nouveau driver in el7 most likely does not support that newer generation card.
NVidia's proprietary driver does.

You do have the 'elrepo' repository.
Install package nvidia-detect.
Then run command nvidia-detect
It should output:

Code: Select all

kmod-nvidia

If it does, then run yum install kmod-nvidia and (if success) reboot.

The driver functions only with the el7's 3.10.0-1160 kernels.
An open question is whether install succeeds you are running the kernel-ml kernel.

I would reboot the machine to the 3.10.0-1160-* kernel, ssh into machine from another machine,
and then install the driver packages.


PS. I would also consider installing a newer Enterprise Linux distro. For example, AlmaLinux 9 or Rocky Linux 9.
The CentOS 7 has EOL in June 2024. It is now nine years old. Very much on "last legs".

[TrevorH]

Also the built in nouveau driver on an EL9 system would probably support that video chipset out of the box...

[almondon]

Thanks @jlehtone and @TrevorH

And sorry for being late.
Yes, in el7, I have nvidia-detect and I receive kmod-nvidia after running nvidia-detect.

Also when I run yum install kmod-nvidia, I receive:
Loaded plugins: fastestmirror, langpacks
You need to be root to perform this command.

Just I could not understand this part:
"I would reboot the machine to the 3.10.0-1160-* kernel, ssh into machine from another machine,
and then install the driver packages."

I reboot to the same el7.
And try to ssh from different machine to the current one.
which driver packages I need to install through ssh?

Also, can I install EL9 kernel?

Cheers

[almondon]

BTW, I think for kernel el9 I need to install centos9.
I'm trying to avoid new OS, as many packages I installed,
and I hope I can use it for one more year.

Cheers

[TrevorH]

You need to be root to perform this command.

Fairly self-explanatory... you were not root so you cannot do this.. Be root and you can.

You can either use a newer kernel OR you can use kmod-nvidia. To use kmod-nvidia you MUST use the CentOS 7 distro kernel.

Also, can I install EL9 kernel?

You can on an EL9 system. You cannot on an EL7 system.

[jlehtone]

Just I could not understand this part:
"I would reboot the machine to the 3.10.0-1160-* kernel, ssh into machine from another machine,
and then install the driver packages."

I reboot to the same el7.
And try to ssh from different machine to the current one.
which driver packages I need to install through ssh?

I did initially take the issue as: "black screen, no display". The use of computer without a display is a bit awkward, isn't it?
(Although, ages ago, I did start and shutdown an Windows NT server successfully while it had no physical display.)
Whenever I do have the "no display" or "stuck GUI" with Linux, I try to ssh in from elsewhere, from system that is not stuck and has display.
A shell session in a machine is a shell session whether started locally or via ssh.

However, you have managed to install kernel-ml, which you must have done while the el7 kernel was in use, so you must have had some display.
Text-mode console, I presume. If you have that, then you don't need ssh.

The el7 kernels do have version 3.10.0-1160-*. The kernel-ml is entirely different version.
The ELRepo's kmod-nvidia is for the 3.10.0-1160-* kernels and you better run one of those while installing it.

[almondon]

Thanks @TrevorH
Sorry I forgot to use root and I did not look carefully at the error message:

I tried as root and I received this:
Error: nvidia-x11-drv conflicts with ocl-icd-2.2.12-1.el7.x86_64
You could try using --skip-broken to work around the problem
You could try running: rpm -Va --nofiles --nodigest

Also I tried as
sudo yum install kmod-nvidia --skip-broken
and it was not successful.

Cheers