I recently rebooted my system running centOS 7.
Post reboot nvidia-smi didn’t work, it gave the error:
NVIDIA-SMI has failed because it couldn't communicate with the NVIDIA driver. Make sure that the latest NVIDIA driver is installed and running.
I have checked the GPUs are still connected using lspci.
The dkms status says this:
nvidia/515.48.07, 3.10.0-1160.66.1.el7.x86_64, x86_64: installed
nvidia/515.48.07, 3.10.0-1160.71.1.el7.x86_64, x86_64: installed
I do not have a blacklist_nvidia file in modprobe.
I tried reinstalling the drivers using the (official documentation) but it said it’s already installed.
I have attached the nvidia-bug-report.log generated on running nvidia-bug-report.sh.
I couldn’t understand the log file fully, but the last line says:
Error: nvmlInit(): Driver Not Loaded
I am not sure what is missing or what I need to do to load the drivers.
Thanks in advance
nvidia-bug-report.log (718.7 KB)
nvidia: Unknown parameter `“NVreg_RestrictProfilingToAdminUsers’
Please remove from module parameters
Thank you!
Yes removing that and rebooting totally enabled the nvidia drivers.
Adding some more info for other users:
The reason I did that was to allow profiling to all users, the documentation says (here) to add a .conf file containing 'options nvidia "NVreg_RestrictProfilingToAdminUsers=0"' in the /etc/modprob.d directory.
I took it literally and added the line (with the single quotes) in a conf file which probably caused the error.
On removing those single quotes and rebooting it totally worked.
So, you’re having trouble with your Nvidia-driver on Centos 7, huh? That’s a bummer, but don’t worry, I’ve got you covered! Have you considered using Extended Lifecycle Support (ELS)? It’s a way to get long-term support for your distro, even if it’s an older version like Centos 7. With ELS, you can get security updates and bug fixes for up to 5 years after the end of standard support. If you’re interested, you can check out https://tuxcare.com - they’re a bunch of Linux experts who specialize in ELS. They can help you get your Nvidia-driver up and running, and answer any other questions you have about ELS.