Kubernetes GPU 环境 （k8s 1.17 docker 18.09 ）

# 添加官方密钥
# docker
curl -fsSL https://download.docker.com/linux/ubuntu/gpg | sudo apt-key add -
# nvidia-docker2
curl -s -L https://nvidia.github.io/nvidia-docker/gpgkey |  sudo apt-key add -
# k8s
curl https://mirrors.aliyun.com/kubernetes/apt/doc/apt-key.gpg | apt-key add -

# 添加仓库
# dokcer 
add-apt-repository "deb [arch=amd64] https://download.docker.com/linux/ubuntu $(lsb_release -cs) stable"
# nvidia-docker2
distribution=$(. /etc/os-release;echo $ID$VERSION_ID)
curl -s -L https://nvidia.github.io/nvidia-docker/$distribution/nvidia-docker.list | sudo tee /etc/apt/sources.list.d/nvidia-docker.list
# k8s
cat <<EOF >/etc/apt/sources.list.d/kubernetes.list
deb https://mirrors.aliyun.com/kubernetes/apt/ kubernetes-xenial main
EOF

# 更新源
apt-get update

# 安装docker
# k8s 1.16 1.17版本还不支持docker 19.03，这里docker安装18.09版本，nvidia-docker2安装对应版本
apt-get install docker-ce=5:18.09.9~3-0~ubuntu-xenial

# 安装nvidia-docker2
apt-get install nvidia-docker2

# 安装k8s
apt-get install -y kubelet kubeadm kubectl
apt-mark hold kubelet kubeadm kubectl

# k8s镜像手动下载
# 国外源的问题，直接把镜像先手动下载
# 使用Azure中国提供的国内源，使用silenceshell提供的脚本
# 不用的话你就自己手动改标签
git clone https://github.com/silenceshell/docker-wrapper.git 
cd docker-wrapper
chmod 777 docker_wrapper.py

./docker_wrapper.py pull  k8s.gcr.io/kube-apiserver:v1.17.0
./docker_wrapper.py pull  k8s.gcr.io/kube-controller-manager:v1.17.0
./docker_wrapper.py pull  k8s.gcr.io/kube-scheduler:v1.17.0
./docker_wrapper.py pull  k8s.gcr.io/kube-proxy:v1.17.0
./docker_wrapper.py pull  k8s.gcr.io/pause:3.1
./docker_wrapper.py pull k8s.gcr.io/etcd:3.4.3-0
./docker_wrapper.py pull  k8s.gcr.io/coredns:1.6.5

# 初始化k8s
# 记住之前 swapoff -a 和 swap挂载去掉
kubeadm init --pod-network-cidr=10.244.0.0/16
mkdir -p "${HOME_DIR}/.kube"
cp /etc/kubernetes/admin.conf "${HOME_DIR}/.kube/config"
chown -R "$(id -u):$(id -g)" "${HOME_DIR}/.kube"

# 初始化网络和gpu plugin
kubectl apply -f kube-flannel.yml
kubectl apply -f nvidia-1.9.10.yml

plugin 和 flannel.zip