Ask the experts，I would like to ask about the issue of not being able to install the NVIDIA 570.86 driver on OPENSUSE 15.5.
Recently, I deployed offline AI on a server with two Tesla V100 32G GPUs, and the operating system is OPENSUSE 15.5. I am currently facing issues installing the driver. According to the NV official website, the V100 supports CUDA 12.8, and there are corresponding driver versions available for download for the operating system. However, whether I download the CUDA RUN file (size: 5G) or the RPM Data Center Driver for Linux x64 570.86.15 | Linux 64-bit, I cannot install them properly. When trying to install with CUDA, the log feedback.
[INFO]: Driver not installed.
[INFO]: Checking compiler version…
[INFO]: gcc location: /usr/bin/gcc
[INFO]: gcc version: gcc version 7.5.0 (SUSE Linux)
[INFO]: Initializing menu
[INFO]: nvidia-fs.setKOVersion(2.24.2)
[INFO]: Setup complete
[INFO]: Installing: Driver
[INFO]: Installing: 570.86.10
[INFO]: Executing NVIDIA-Linux-x86_64-570.86.10.run --ui=none --no-questions --accept-license --disable-nouveau --no-cc-version-check --install-libglvnd 2>&1
[INFO]: Finished with code: 256
[ERROR]: Install of driver component failed. Consult the driver log at /var/log/nvidia-installer.log for more details.
[ERROR]: Install of 570.86.10 failed, quitting
When trying to install with RPM, the feedback is “Header V4 RSA/SH512 Signature, key id xxxxx: nokey.”
Since the device does not have internet access, I downloaded the two NVIDIA online repositories to the local machine. After allowing the device to access the local area network, I could only install driver 470.256 and CUDA 11.8, but I cannot install 12.8. What could be the issue? I have imported both GPG keys from the online repository, but the RPM installation still prompts NOKEY.