anolisos8.8安装显卡及cuda工具
一、目录
1、测试环境
2、安装显卡驱动
3、安装cuda工具
4、配置容器运行时
5、K8S集群安装nvidia插件
二、测试环境
操作系统：Anolis OS 8.8
内核版本：5.10.134-13.an8.x86_64
显卡安装版本：525.147.05
cuda版本：V10.2.89
外网要求：必须
三、安装显卡驱动
3.1、禁用nonveau
shell
[root@localhost ~]# wget https://ops-publicread-1257137142.cos.ap-beijing.myqcloud.com/shell/disable_nouveau.sh
[root@localhost ~]# bash disable_nouveau.sh
[root@localhost ~]# lsmod | grep nouveau
#重启服务器再次进行检测
[root@localhost ~]# reboot
[root@localhost ~]# lsmod | grep nouveau3.2、下载显卡驱动并安装
显卡下载地址：https://www.nvidia.com/Download/Find.aspx?lang=en-us#
注：根据显卡型号选择对应驱动进行下载！
shell
[root@localhost src]# lspci |grep NVIDIA
13:00.0 3D controller: NVIDIA Corporation TU104GL [Tesla T4] (rev a1)
[root@localhost src]# wget https://us.download.nvidia.cn/tesla/525.147.05/NVIDIA-Linux-x86_64-525.147.05.run
[root@localhost src]# bash NVIDIA-Linux-x86_64-525.147.05.run
#根据提示进行安装如下则安装完成！
检测
shell
[root@localhost src]# nvidia-smi
Tue Dec 12 10:16:35 2023
+-----------------------------------------------------------------------------+
| NVIDIA-SMI 525.147.05 Driver Version: 525.147.05 CUDA Version: 12.0 |
|-------------------------------+----------------------+----------------------+
| GPU Name Persistence-M| Bus-Id Disp.A | Volatile Uncorr. ECC |
| Fan Temp Perf Pwr:Usage/Cap| Memory-Usage | GPU-Util Compute M. |
| | | MIG M. |
|===============================+======================+======================|
| 0 Tesla T4 Off | 00000000:13:00.0 Off | 0 |
| N/A 63C P0 30W / 70W | 2MiB / 15360MiB | 5% Default |
| | | N/A |
+-------------------------------+----------------------+----------------------+
+-----------------------------------------------------------------------------+
| Processes: |
| GPU GI CI PID Type Process name GPU Memory |
| ID ID Usage |
|=============================================================================|
| No running processes found |
+-----------------------------------------------------------------------------+四、安装CUDA工具
4.1、官网下载指定版本CUDA
https://developer.nvidia.com/cuda-toolkit-archive
4.2、安装CUDA
shell
[root@localhost src]# wget https://developer.download.nvidia.com/compute/cuda/10.2/Prod/local_installers/cuda_10.2.89_440.33.01_linux.run
[root@localhost src]# sh cuda_10.2.89_440.33.01_linux.run
#加载程序耗时3分钟
---------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------
x End User License Agreement x
x - x
x x
x x
x Preface x
x - x
x x
x The Software License Agreement in Chapter 1 and the Supplement x
x in Chapter 2 contain license terms and conditions that govern x
x the use of NVIDIA software. By accepting this agreement, you x
x agree to comply with all the terms and conditions applicable x
x to the product(s) included herein. x
x x
x x
x NVIDIA Driver x
x x
x x
x Description x
x x
x This package contains the operating system driver and x
xq x
x Do you accept the above EULA? (accept/decline/quit): x
x accept x
---------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------
#输入accept回车
---------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------
x CUDA Installer se Agreement x
x - [ ] Driver x
x [ ] 440.33.01 x
x + [X] CUDA Toolkit 10.2 x
x [X] CUDA Samples 10.2 x
x [X] CUDA Demo Suite 10.2 x
x [X] CUDA Documentation 10.2 x
x Options x
x Install x
x x
x x
x x
x x
x x
x VIDIA Driver x
x x
x x
x escription x
x x
x x
x
x x
x Up/Down: Move | Left/Right: Expand | 'Enter': Select | 'A': Advanced options x
---------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------
#去掉显卡驱动选择install继续4.3、设置cuda环境变量
shell
[root@localhost ~]# echo "export PATH=/usr/local/cuda/bin:$PATH" >> /etc/profile
[root@localhost ~]# echo "export LD_LIBRARY_PATH=/usr/local/cuda/lib64:$LD_LIBRARY_PATH" >> /etc/profile
[root@localhost ~]# source /etc/profile
[root@localhost ~]# nvcc --version
nvcc: NVIDIA (R) Cuda compiler driver
Copyright (c) 2005-2019 NVIDIA Corporation
Built on Wed_Oct_23_19:24:38_PDT_2019
Cuda compilation tools, release 10.2, V10.2.89五、配置容器运行时
5.1、安装显卡容器运行时
shell
#添加阿里docker-ce源
# step 1: 安装必要的一些系统工具
[root@localhost ~]# yum install -y yum-utils device-mapper-persistent-data lvm2
# Step 2: 添加软件源信息
[root@localhost ~]# yum-config-manager --add-repo https://mirrors.aliyun.com/docker-ce/linux/centos/docker-ce.repo
# Step 3
[root@localhost ~]# sed -i 's+download.docker.com+mirrors.aliyun.com/docker-ce+' /etc/yum.repos.d/docker-ce.repo
# Step 4: 更新并安装Docker-CE
[root@localhost ~]# yum makecache
# Step 5: 安装显卡容器运行时
[root@localhost ~]# yum -y install nvidia-docker25.2、配置containerd支持显卡
shell
# Step1 : 安装containerd
[root@localhost ~]# yum -y install containerd.io
# Step2 ：生成默认配置
[root@localhost ~]# containerd config default > /etc/containerd/config.toml
# Step3 ：修改containerd配置 /etc/containerd/config.toml
###############################################################
...
[plugins."io.containerd.grpc.v1.cri".containerd]
snapshotter = "overlayfs"
default_runtime_name = "runc"
no_pivot = false
...
[plugins."io.containerd.grpc.v1.cri".containerd.runtimes]
[plugins."io.containerd.grpc.v1.cri".containerd.runtimes.runc]
runtime_type = "io.containerd.runtime.v1.linux" # 将此处 runtime_type 的值改成 io.containerd.runtime.v1.linux
...
[plugins."io.containerd.runtime.v1.linux"]
shim = "containerd-shim"
runtime = "nvidia-container-runtime" # 将此处 runtime 的值改成 nvidia-container-runtime
...
###########################################################
# Step4 : 启动containerd
[root@localhost ~]# systemctl start containerd && systemctl enable containerd
# Step5 : 跑测试容器测试
[root@localhost ~]# ctr image pull docker.io/nvidia/cuda:11.2.2-base-ubuntu20.04
[root@localhost ~]# ctr run --rm -t \
> --runc-binary=/usr/bin/nvidia-container-runtime \
> --env NVIDIA_VISIBLE_DEVICES=all \
> docker.io/nvidia/cuda:11.2.2-base-ubuntu20.04 \
> cuda-11.6.2-base-ubuntu20.04 nvidia-smi
Tue Dec 12 03:01:10 2023
+-----------------------------------------------------------------------------+
| NVIDIA-SMI 525.147.05 Driver Version: 525.147.05 CUDA Version: 12.0 |
|-------------------------------+----------------------+----------------------+
| GPU Name Persistence-M| Bus-Id Disp.A | Volatile Uncorr. ECC |
| Fan Temp Perf Pwr:Usage/Cap| Memory-Usage | GPU-Util Compute M. |
| | | MIG M. |
|===============================+======================+======================|
| 0 Tesla T4 Off | 00000000:13:00.0 Off | 0 |
| N/A 66C P0 30W / 70W | 2MiB / 15360MiB | 4% Default |
| | | N/A |
+-------------------------------+----------------------+----------------------+
+-----------------------------------------------------------------------------+
| Processes: |
| GPU GI CI PID Type Process name GPU Memory |
| ID ID Usage |
|=============================================================================|
| No running processes found |
+-----------------------------------------------------------------------------+5.3、配置Docker支持显卡
shell
# Step1 : 安装docker
[root@localhost ~]# yum install docker-ce-23.0.6 -y
# Step2 : 配置docker容器运行时，并启动docker
#修改cgroup驱动为systemd[k8s官方推荐]、限制容器日志量、修改存储类型
[root@localhost ~]# mkdir /etc/docker -p
[root@localhost ~]# cat > /etc/docker/daemon.json <<EOF
{
"registry-mirrors": [
"https://tf72mndn.mirror.aliyuncs.com"
],
"exec-opts": ["native.cgroupdriver=systemd"],
"log-opts": {
"max-file": "3",
"max-size": "500m"
},
"runtimes": {
"nvidia": {
"path": "/usr/bin/nvidia-container-runtime",
"runtimeArgs": []
}
}
}
EOF
[root@localhost ~]# systemctl daemon-reload
[root@localhost ~]# systemctl restart docker
[root@localhost ~]# systemctl enable docker
# Step3 : 启动docker测试容器
[root@localhost ~]# docker run --runtime=nvidia --rm nvidia/cuda:11.0-base nvidia-smi
Unable to find image 'nvidia/cuda:11.0-base' locally
11.0-base: Pulling from nvidia/cuda
54ee1f796a1e: Pull complete
f7bfea53ad12: Pull complete
46d371e02073: Pull complete
b66c17bbf772: Pull complete
3642f1a6dfb3: Pull complete
e5ce55b8b4b9: Pull complete
155bc0332b0a: Pull complete
Digest: sha256:774ca3d612de15213102c2dbbba55df44dc5cf9870ca2be6c6e9c627fa63d67a
Status: Downloaded newer image for nvidia/cuda:11.0-base
Tue Dec 12 03:10:32 2023
+-----------------------------------------------------------------------------+
| NVIDIA-SMI 525.147.05 Driver Version: 525.147.05 CUDA Version: 12.0 |
|-------------------------------+----------------------+----------------------+
| GPU Name Persistence-M| Bus-Id Disp.A | Volatile Uncorr. ECC |
| Fan Temp Perf Pwr:Usage/Cap| Memory-Usage | GPU-Util Compute M. |
| | | MIG M. |
|===============================+======================+======================|
| 0 Tesla T4 Off | 00000000:13:00.0 Off | 0 |
| N/A 64C P0 30W / 70W | 2MiB / 15360MiB | 5% Default |
| | | N/A |
+-------------------------------+----------------------+----------------------+
+-----------------------------------------------------------------------------+
| Processes: |
| GPU GI CI PID Type Process name GPU Memory |
| ID ID Usage |
|=============================================================================|
| No running processes found |
+-----------------------------------------------------------------------------+六、K8S集群安装nvidia插件
shell
# Step1 : GPU主机打标签
[root@localhost ~]# kubectl label node node9 nvidia.com=gpu
# Step2 : K8S集群安装GPU驱动插件（仅需要安装一次！）
[root@localhost ~]# kubectl apply -f https://raw.githubusercontent.com/NVIDIA/k8s-device-plugin/1.0.0-beta4/nvidia-device-plugin.yml
# Step3 : 带GPU资源主机GPU资源信息
[root@localhost ~]# kubectl describe node node9 |grep gpu
gpu/type=nvidia
nvidia.com/gpu: 1
nvidia.com/gpu: 1
nvidia.com/gpu 0 0
# Step4 : 部署使用GPU资源测试容器
apiVersion: v1
kind: Pod
metadata:
name: cuda-vector-add
spec:
restartPolicy: OnFailure
containers:
- name: cuda-vector-add
#image: "k8s.gcr.io/cuda-vector-add:v0.1"
image: "docker.io/nvidia/cuda:11.0.3-base-ubuntu20.04"
command:
- nvidia-smi
resources:
limits:
nvidia.com/gpu: 1其他：disable_nouveau.sh 脚本内容
#!/bin/bash
echo -e "\033[32m>>>>>>>>更新系统内核，请耐心等待！\033[0m"
yum -y install gcc make elfutils-libelf-devel libglvnd-devel kernel-devel epel-release
yum -y install dkms
rm -f /etc/modprobe.d/blacklist-nvidia-nouveau.conf /etc/modprobe.d/nvidia-unsupported-gpu.conf
echo blacklist nouveau | tee /etc/modprobe.d/blacklist-nvidia-nouveau.conf && \
echo options nouveau modeset=0 | tee -a /etc/modprobe.d/blacklist-nvidia-nouveau.conf && \
echo options nvidia NVreg_OpenRmEnableUnsupportedGpus=1 | tee /etc/modprobe.d/nvidia-unsupported-gpu.conf
mv /boot/initramfs-$(uname -r).img /boot/initramfs-$(uname -r)-nouveau.img
dracut /boot/initramfs-$(uname -r).img $(uname -r)