Black screen after install of nvidia driver ubuntu

nvidia_developer_85 · July 5, 2021, 1:08pm

Hi! I have a dual boot on a desktop PC (Ubuntu 20.04 and Windows 10), I have an ASUS motherboard with an AMD Ryzen processor and an NVIDIA GeForce RTX 3080. The PC stopped being able to boot after I downloaded updates: “sudo apt update” and “sudo apt upgrade”. Currently, after attempting a normal boot of Ubuntu 20.04, I get a blank screen with a terminal cursor blinking and then the terminal cursor is frozen, the PC remains unresponsive. Currently I can only boot the computer in {recovery mode + network + root access}.

I tried reinstalling the nvidia drivers (versions 460 and 465) several times, and have tried different proposed solutions listed on this and other threads.

Here is my nvidia-bug-report.log
nvidia-bug-report.log.gz (99.6 KB)

I would appreciate any solution or something that I could test next. Thanks in advance!

nvidia_developer_85 · July 5, 2021, 2:15pm

Actually, for anybody having this issue, I just googled a few times more. I found that there might be some issue in the nvidia-driver-460 after the version 80. Solution, install by hand the version 460.73. (This solution worked for me.)

I got the tip from here:
issue: Ubuntu freezing on boot after Nvidia driver installation
solution: 465.24.02 page fault - #102 by chainofflowers
driver version that worked for me (460.73): https://us.download.nvidia.com/XFree86/Linux-x86_64/460.73.01/NVIDIA-Linux-x86_64-460.73.01.run

abanoub.ghobrial · July 7, 2021, 8:28am

Hi, thanks for your reply. Here is my nvidia log file attached. Thanks :)
nvidia-bug-report.log.gz (1.2 MB)

generix · July 7, 2021, 8:56am

You have a hybrid graphics notebook so both intel and nvidia gpu will have to work.
The kernel is too old for the intel gpu, please upgrade HWE:
https://wiki.ubuntu.com/Kernel/LTSEnablementStack
Then nouveau is still active, please blacklist it, create /etc/modprobe.d/nouveau-blacklist.conf

blacklist nouveau

and run
sudo update-initramfs -u
Finally, please delete /etc/X11/xorg.conf

andreas.b.scharf1 · July 7, 2021, 7:01pm

Hi everybody,
I have freshly installed Ubuntu 20.04 in a dual boot setup with WIndows10 (also freshly installed, and Nvidia driver is working). Installing the newest Nvidia driver 460.80 on Ubuntu results in black screen after boot with my default kernel 5.8.0-59 (using Ctr+Alt+F3 for a terminal to get some info). However I can boot the backup kernel 5.8.0.43. But in both cases nvidia-smi reports: “NVIDIA-SMI has failed because it couldn’t communicate with the NVIDIA driver. Make sure that the latest NVIDIA driver is installed and running.” Now I tried some solutions presented here and others in the Nvidia forum. None worked. The bug-report is attached but please note that a black screen suddenly showed up during the run of “nvidia-bug-report.sh” … resulting in hitting the reset button on my machine.
Thank you for your support.
nvidia-bug-report.log.gz (146.4 KB)

yangyiwei2000 · July 10, 2021, 9:15pm

nvidia-bug-report.log.gz (342.3 KB)
Hello, I’m using AERO15-XA 9750h+2070M. I was installing 20.04 with 460+cuda10.1 and works fine, but when I upgrade it to 470 and install cuda11, it goes black screen with only cursor afterward. I switch between 460, 460-server, and 470, and doesn’t work. here’s my bug report. Plz, help.

purplebeehive · July 13, 2021, 8:19am

Hi!
I have a similar issue.
I installed the recommended, tested nvidia driver
460 on my Ryzen 5600H, gtx 1650 laptop and it won’t reboot. Black screen, blinking cursor. I don’t know how to remove it (I’ve read a lot of forums but I’m scared I’ll mess something up) and which Nvidia version should I switch to (or if I should give up for now and go back to nouveau).
Thank you very much!

Edit: forgot to mention this was after a ubuntu 21.04 fresh install (worked fine) and switching to proprietary drivers (crashed).

flow · July 15, 2021, 7:21am

I was trying to run Ubuntu 20.04 on an Alienware m15 r4 with RTX 3080 and had the same issue with the internal display not working (external display was working). For me upgrading to the 470 beta driver solved the issue.

johnson9510 · August 23, 2021, 6:26pm

I upgraded from 16.04 with an install of 396 to 20.04. The driver installer no longer worked on this version so I tried to install the latest driver of nvidia-driver-470. After doing so I end up booting into a black screen, the log is attached. nvidia-driver-460 also doesn’t work. The only version that is working so far has been nvidia-driver-390, but I would like to upgrade to a newer driver to use newer cuda versions.

nvidia-bug-report.log (354.5 KB)

happykoder · August 25, 2021, 4:49am

made an account to paste what I had to get this trashfire working on a new thinkpad. why is this so much suffering

thinkpad extreme g3 32g ram
nvidia 1650 ti

bios settings that worked:

graphics mode: discrete
secure boot off
slow boot

os settings:

ubuntu 21.04 (18, 20 didnt work)
kernel 5.11.0-31 generic (5.10 might be easier to get working but I didn’t have any packages for it, avoid oem kernels)
driver: nvidia-driver-450 (and related deps)
nvidia-prime: yes
blacklist nouveau: yes

grub

GRUB_CMDLINE_LINUX_DEFAULT=“text”
GRUB_TERMINAL=console

Tried so many installs of the nvidia driver manually and so many versions/kernels, they all resulted in some form of a black screen/freeze. after may kernel swaps, reinstalls I tried an older ubuntu driver meta-package (nvidia-driver-450) and for some reason I haven’t had any black screens. I’m probably 1 apt-get install from calamity.

anyways, good luck friends

(to anyone reading from ubuntu or nvidia, I will pay you so much money for stable, easy to scaffold linux installs for portable desktop computing. why is this still so hard in 2021? )

ubuntero · August 27, 2021, 11:51am

EDIT: I have managed to install video drivers for both Intel and Nvidia on MX Linux 19.4 KDE Plasma, AHS version.

I have prepared fresh Linux Mint 20.2 install on HP Pavilion 15-eg0079nw equipped with both: Intel i915 and Nvidia GeForce MX450 GPUs, 11th generation processor.

After installing updates, I installed nvidia v 460.91.03 driver from distro repositories. After rebooting I received black screen, but managed to ssh and created log in attachment. More info on Mint Forum

Can you help resolve the issue?

fbcanyagmur · August 28, 2021, 12:16am

Guys can you help me with this ?

rojas70 · September 10, 2021, 10:42am

Greetings,

I am have long been struggling to get rid of the black screen.
Ubuntu 20.04, Nvidia-driver-470
GM107GLM [Quadro M2000M]

Have tried all solutions above many times but have not succeeded in months. With blackscreen I can still login via terminal and see all specs with nvidia-smi…

To start the gui I load older kernels like those in the 5.4.0-5XX-generic family to activate the gui. But at that stage I get the

NVIDIA-SMI has failed because it couldn’t communicate with the NVIDIA driver. Make sure that the latest NVIDIA driver is installed and running.

Attached is my log report.

Please let me know if you can suggest any fix.
Thank you

nvidia-bug-report.log (124.8 KB)

ioguiman · September 14, 2021, 4:09pm

@generix I have a fresh ubuntu installation with this very same problem as posted here:

Considering that this is a fresh installation I don’t have the older kernel. Is it possible to downgrade the kernel in this situation?

hinchliff · September 24, 2021, 11:28pm

I’m seeing the same issues, but also Screen Flickering which I did not see above – so I created a new top-level post here: Black Screen and Screen Flickering in Ubuntu 20.04 with 470 drivers

xuanyingchan6 · October 12, 2021, 4:55pm

I just upgraded my OS from ubuntu 20.04 to 21.04, and I found that my sceen stuck at

/dea/sda3:clean XXXX/XXXX files, XXXX/XXXX blocks

I have tried all methods mentioned above, incuding use ,run file and rep to install the driver. After a loooooooooong time trial and error, I fixed the problem by modifying my Xwrapper.config

Here are the details:

When I run sudo apt purge nvidia* to remove the installed driver, the system can work smoothly. As long as I put the driver in, the system stuck, but can use Ctrl + Alt + F2 to use tty.

And I happend to run startx, the error turned out

I just removed the line need_root_rights: yes, then startx and the desktop showed up.

My driver version:

Driver Version: 470.63.01

My kenerl version:

5.11.0-37-generic

Hope this can help someone.

JCambeiro · November 30, 2021, 9:22am

Dear all,

I write here after reading and trying all the related threads I could find about this.
I have an HPE server Proliant DL380 Gen10 with integrated graphics card (MATROX MGA G200eH3), with Ubuntu 20.04, and an nvidia Quadro RTX4000 (we’ve tested an RTX6000 as well).
Installed nvidia 470 drivers, no issues detected during installation

Problems encounteres so far:

xrandr command doesn’t ever show the displayport outputs of the nvidia card, only can see the integrated VGA output.
when using nvidia-smi, the GPU load is always at 0%. we tried with software which we know in other servers is succesfully using the GPU, but in this server it all goes to CPU.
we had several black screen problems, but we were able to sort it out by deleting the file /etc/X11/xorg.conf (fix that we saw in other thread). Before this, we had to purge the drivers for the server to restart.
All nomodeset parameters found in different files were deleted.
Server is NOT in secure boot mode.
No stray blacklist files were found ( /lib/modprobe.d/blacklist-nvidia.conf etc/modprobe.d/blacklist-nvidia.conf)
rest of the commands seem to show a “no problem” view: ‘lshw -c video’ shows both cards, nvidia-settings seem to work fine, prime-select query always shows ‘nvidia’, lspci show the nvidia modules…

I attach some of the outputs here:

innovation@innovation-holographics:/usr/sbin$ sudo lshw -c video

*-display
description: VGA compatible controller
product: MGA G200eH3
vendor: Matrox Electronics Systems Ltd.
physical id: 0.1
bus info: pci@0000:01:00.1
version: 02
width: 32 bits
clock: 33MHz
capabilities: pm msi pciexpress vga_controller bus_master cap_list rom
configuration: driver=mgag200 latency=0
resources: irq:17 memory:d8000000-d8ffffff memory:d9b98000-d9b9bfff memory:d9000000-d97fffff memory:c0000-dffff
*-display
description: VGA compatible controller
product: TU104GL [Quadro RTX 4000]
vendor: NVIDIA Corporation
physical id: 0
bus info: pci@0000:86:00.0
version: a1
width: 64 bits
clock: 33MHz
capabilities: pm msi pciexpress vga_controller bus_master cap_list rom
configuration: driver=nvidia latency=0
resources: iomemory:d7f0-d7ef iomemory:d7f0-d7ef irq:433 memory:f0000000-f0ffffff memory:d7fe0000000-d7fefffffff memory:d7ff0000000-d7ff1ffffff ioport:a000(size=128) memory:f1080000-f10fffff

innovation@innovation-holographics:/usr/sbin$ lspci -k |grep nvidia
Kernel driver in use: nvidia
Kernel modules: nvidiafb, nouveau, nvidia_drm, nvidia
Kernel driver in use: nvidia-gpu
Kernel modules: i2c_nvidia_gpu

±----------------------------------------------------------------------------+
| Processes: |
| GPU GI CI PID Type Process name GPU Memory |
| ID ID Usage |
|=============================================================================|
| 0 N/A N/A 2383 G /usr/lib/xorg/Xorg 4MiB |
| 0 N/A N/A 3332 G /usr/lib/xorg/Xorg 4MiB |
±----------------------------------------------------------------------------+

innovation@innovation-holographics:/usr/sbin$ xrandr
Screen 0: minimum 320 x 200, current 1920 x 1080, maximum 4096 x 4096
VGA-1 connected primary 1920x1080+0+0 (normal left inverted right x axis y axis) 596mm x 335mm
1600x900 60.00
1280x1024 75.02 60.02
1152x864 75.00
1024x768 75.03 60.00
800x600 75.00 60.32
640x480 75.00 59.94
1920x1080_60.00 59.96*

innovation@innovation-holographics:/usr/sbin$ prime-select query
nvidia

Thanks for any help, and sorry if this an obvious problem, but we’ve tested everything we could find in these threads, with no success…

generix · December 3, 2021, 9:02pm

The matrox driver is a very simple frambuffer driver which doesn’t support any PRIME functions. You can only connect the monitor to the nvidia and configure xorg to use it as primary gpu.