[Starlingx-discuss] If starlingx 8.0 support Nvidia gpu-operator
Hi Stalingx Community, I'm using the starlingx 8.0. I tried to install gpu-operator v22.9.1 version helm chart but failed with the fllowing error message: ``` ========== NVIDIA Software Installer ========== Starting installation of NVIDIA driver version 525.60.13 for Linux kernel version 5.10.0-6-amd64 Stopping NVIDIA persistence daemon... Unloading NVIDIA driver kernel modules... Unmounting NVIDIA driver rootfs... Checking NVIDIA driver packages... Updating the package cache... Resolving Linux kernel version... Stopping NVIDIA persistence daemon... Unloading NVIDIA driver kernel modules... Unmounting NVIDIA driver rootfs... Could not resolve Linux kernel version ``` and I also tried to install the nvidia dirver manually with the 'sudo ./NVIDIA-Linux-x86_64-525.60.13.run' but it let me to install gcc. when I tried to install the gcc pacage, i got the error "dpkg: error: unable to access the dpkg database directory /var/lib/dpkg: Read-only file system". So I want to know if stx8.0 support gpu-operator? and how I can install it sucessfully? Thank you for your help in advance. Best regards,Aqan
Hi Aqan, The gpu-operator is not integrated with StarlingX as a system fluxcd application which is the typical model for supporting various containerized applications/projects in StarlingX. There was an attempt at a loose integration of an older version of the gpu-operator (1.8.1) by packaging the helm chart only, but that is no longer maintained for stx.8.0 as per this commit: https://opendev.org/starlingx/integ/commit/575ba4251526cc34162d36c0e3a0ba5be... Regarding the error you are seeing when trying to build/install packages directly, this is tied to the underlying ostree setup in stx.8.0 which makes the filesystem read-only. I believe there is a workaround to unlock the filesystem for debugging/experimenting. @Tao, Yue<mailto:Yue.Tao@windriver.com> can the distro-os team provide some further guidance on this? Regards, Ghada From: aqan_bj@sina.com <aqan_bj@sina.com> Sent: Wednesday, September 6, 2023 11:27 PM To: starlingx-discuss <starlingx-discuss@lists.starlingx.io> Subject: [Starlingx-discuss] If starlingx 8.0 support Nvidia gpu-operator CAUTION: This email comes from a non Wind River email account! Do not click links or open attachments unless you recognize the sender and know the content is safe. Hi Stalingx Community, I'm using the starlingx 8.0. I tried to install gpu-operator v22.9.1 version helm chart but failed with the fllowing error message: ``` ========== NVIDIA Software Installer ========== Starting installation of NVIDIA driver version 525.60.13 for Linux kernel version 5.10.0-6-amd64 Stopping NVIDIA persistence daemon... Unloading NVIDIA driver kernel modules... Unmounting NVIDIA driver rootfs... Checking NVIDIA driver packages... Updating the package cache... Resolving Linux kernel version... Stopping NVIDIA persistence daemon... Unloading NVIDIA driver kernel modules... Unmounting NVIDIA driver rootfs... Could not resolve Linux kernel version ``` and I also tried to install the nvidia dirver manually with the 'sudo ./NVIDIA-Linux-x86_64-525.60.13.run<https://urldefense.com/v3/__http:/NVIDIA-Linux-x86_64-525.60.13.run__;!!AjveYdw8EvQ!YYTSkiBKVlfmXDpvZG4knt6h7Xl8FNkUP6fVYrKG_g-Ltn7VDxP-GEy35LnlWRARu6sS-KeW52dtoN6FyqXl3g$>' but it let me to install gcc. when I tried to install the gcc pacage, i got the error "dpkg: error: unable to access the dpkg database directory /var/lib/dpkg: Read-only file system". So I want to know if stx8.0 support gpu-operator? and how I can install it sucessfully? Thank you for your help in advance. Best regards, Aqan
participants (2)
-
aqan_bj@sina.com
-
Khalil, Ghada