OMV安装NVIDIA驱动供Docker下Jellyfin使用
服务器硬件更新了一轮，顺便搞了块1050Ti给PVE直通OMV的Jellyfin做硬解，皮衣黄的驱动安装也是费了一番功夫，安装方法有两种，下文逐一介绍。
小记：OMV每次系统更新时有可能会影响驱动，需要按照步骤重新安装一遍驱动，才能正常启动Jellyfin。
方法一：通过软件源直接安装
移除已安装的驱动套件
sudo apt autoremove cuda* nvidia* --purge添加闭源软件包
sudo apt install software-properties-common -yBookworm （Debian12）以上版本用:
sudo add-apt-repository contrib non-free-firmwareBullseye and Buster (Debian11\10）以下版本用:
sudo add-apt-repository contrib non-free更新
sudo apt update安装依赖
sudo apt install build-essential gcc dirmngr ca-certificates software-properties-common apt-transport-https dkms curl -y选择对应Debian版本导入key
Bookworm, run:
curl -fSsL https://developer.download.nvidia.com/compute/cuda/repos/debian12/x86_64/3bf863cc.pub | sudo gpg --dearmor | sudo tee /usr/share/keyrings/nvidia-drivers.gpg > /dev/null 2>&1Bullseye, run:
curl -fSsL https://developer.download.nvidia.com/compute/cuda/repos/debian11/x86_64/3bf863cc.pub | sudo gpg --dearmor | sudo tee /usr/share/keyrings/nvidia-drivers.gpg > /dev/null 2>&1Buster, use:
curl -fSsL https://developer.download.nvidia.com/compute/cuda/repos/debian10/x86_64/3bf863cc.pub | sudo gpg --dearmor | sudo tee /usr/share/keyrings/nvidia-drivers.gpg > /dev/null 2>&1添加英伟达的仓库源
Bookworm, use:
echo 'deb [signed-by=/usr/share/keyrings/nvidia-drivers.gpg] https://developer.download.nvidia.com/compute/cuda/repos/debian12/x86_64/ /' | sudo tee /etc/apt/sources.list.d/nvidia-drivers.listBullseye, use:
echo 'deb [signed-by=/usr/share/keyrings/nvidia-drivers.gpg] https://developer.download.nvidia.com/compute/cuda/repos/debian11/x86_64/ /' | sudo tee /etc/apt/sources.list.d/nvidia-drivers.listBuster, use:
echo 'deb [signed-by=/usr/share/keyrings/nvidia-drivers.gpg] https://developer.download.nvidia.com/compute/cuda/repos/debian10/x86_64/ /' | sudo tee /etc/apt/sources.list.d/nvidia-drivers.list更新
sudo apt update安装CUDA和NVIDIA驱动
列表看你的系统支持到哪个版本
apt show cuda-drivers -a截止发文是555版本
sudo apt install nvidia-driver cuda-drivers-555 cuda -y方法二：到英伟达官网下载安装包进行安装
如果软件源下载网络不好，可以尝试这种方法，本方法需要先关闭安全启动，参见安装完毕后下文
https://www.nvidia.cn/drivers/lookup/
进入官网搜索所用显卡的驱动，搜索之后会列出全部受支持的驱动版本，可以选择下载【推荐驱动】，或者有特殊需求的也可以尝试最新的测试版，下载完成后上传至服务器。
sudo sh NVIDIA-Linux-x86_64-555.58.02.run编译时如果提示要禁用nouveau
vim /etc/modprobe.d/blacklist.conf
# 最后一行添加
blacklist nouveau
# 更新、重启
sudo update-initramfs -u
reboot
lsmod | grep nouveau #无返回结果说明禁用成功提示unable to find the kernel source tree for the currently running kernel
sudo apt install linux-headers-$(uname -r)https://developer.nvidia.com/cuda-toolkit-archive
同样进入官网搜索需要的CUDA，如果NVIDIA驱动选择的比较旧，可以查一下版本对应的CUDA版本，下载完成后上传至服务器，依照CUDA页面给的流程操作。
如果你跟我一样是x86_64的debian12，本地安装可以参考如下。
wget https://developer.download.nvidia.com/compute/cuda/12.5.0/local_installers/cuda-repo-debian12-12-5-local_12.5.0-555.42.02-1_amd64.deb
sudo dpkg -i cuda-repo-debian12-12-5-local_12.5.0-555.42.02-1_amd64.deb
sudo cp /var/cuda-repo-debian12-12-5-local/cuda-*-keyring.gpg /usr/share/keyrings/
sudo add-apt-repository contrib
sudo apt-get update
sudo apt-get -y install cuda-toolkit-12-5通过以上任一一种方法安装完毕后
先重启
sudo reboot顺利完成安装，输入下列命令就可以看到英伟达的驱动信息
nvidia-smi如果提示：
NVIDIA-SMI has failed because it couldn't communicate with the NVIDIA driver. Make sure that the latest NVIDIA driver is installed and running.可以尝试关闭安全启动
PVE中的OMV操作：
启动虚拟机，出现Proxmox画面时，按ESC，进入BIOS。
选择【Device Manager】——【Secure Boot Configuration】
取消【Attempt Secure Boot】的【X】如果启动时界面显示基带内核无法加载，可尝试：
apt install linux-headers-amd64 -y给jellyfin使用前准备
安装nvidia-container-toolkit
apt install -y nvidia-container-toolkit配置 Docker 守护进程以识别 NVIDIA 容器运行时
nvidia-ctk runtime configure --runtime=docker重启 docker
systemctl restart dockerjellyfin配置
jellyfin:
image: nyanmisaka/jellyfin
container_name: jellyfin
runtime: nvidia #加
ports:
- "8096:8096"
volumes:
- /fonts:/usr/share/fonts/truetype/dejavu
- /config:/config
- /media:/media
environment:
- PUID=0
- PGID=0
- NVIDIA_VISIBLE_DEVICES=all #加
network_mode: bridge
restart: unless-stopped参考文章：
https://linuxcapable.com/how-to-install-cuda-on-debian-linux/
https://blog.csdn.net/lemonxiaoxiao/article/details/107690494