网站建设地图怎么设置网站创建免费用户
目录
1. 部署cri-docker （所有集群节点）
2. 升级master节点
3. 升级worker节点
4. 部署containerd
1. 部署cri-docker （所有集群节点）
k8s从1.24版本开始移除了dockershim，所以需要安装cri-docker插件才能使用docker
软件下载：GitHub - Mirantis/cri-dockerd: dockerd as a compliant Container Runtime Interface for Kubernetes
安装
rpm -ivh cri-dockerd-0.3.5.20231016182601.cd730ff8-0.el7.x86_64.rpm
配置cri-docker
vim /usr/lib/systemd/system/cri-docker.service[Service]
Type=notify
ExecStart=/usr/bin/cri-dockerd --container-runtime-endpoint fd:// --network-plugin=cni --pod-infra-container-image=reg.westos.org/k8s/pause:3.7systemctl daemon-reload
systemctl enable --now cri-docker
ll /var/run/cri-dockerd.sock
2. 升级master节点
首先上传镜像到harbor仓库，便于升级
docker load -i k8s-v1.24.17.tar
docker images | grep k8s
docker push reg.westos.org/k8s/kube-apiserver:v1.24.17
docker push reg.westos.org/k8s/kube-proxy:v1.24.17
docker push reg.westos.org/k8s/kube-scheduler:v1.24.17
docker push reg.westos.org/k8s/kube-controller-manager:v1.24.17
docker push reg.westos.org/k8s/pause:3.7
升级kubeadm 执行升级
yum install -y kubeadm-1.24.17-0
kubeadm upgrade plan
修改节点套接字
kubectl edit nodes k8s1kubeadm.alpha.kubernetes.io/cri-socket: unix:///var/run/cri-dockerd.sockkubeadm upgrade apply v1.24.17
腾空节点
kubectl drain k8s1 --ignore-daemonsets
升级kubelet
yum install -y kubelet-1.24.17-0 kubectl-1.24.17-0
配置kubelet使用cri-docker
vim /var/lib/kubelet/kubeadm-flags.env
KUBELET_KUBEADM_ARGS="--pod-infra-container-image=reg.westos.org/k8s/pause:3.7 --container-runtime=remote --container-runtime-endpoint=unix:///var/run/cri-dockerd.sock"
重启kubelet
systemctl daemon-reload
systemctl restart kubelet
解除节点保护
kubectl uncordon k8s1
完成升级
3. 升级worker节点
升级kubeadm 执行升级
yum install -y kubeadm-1.24.17-0
kubeadm upgrade node
腾空节点 #需要在master节点执行
kubectl drain k8s2 --ignore-daemonsets
kubectl drain k8s3 --ignore-daemonsets
升级kubelet
yum install -y kubelet-1.24.17-0 kubectl-1.24.17-0
配置kubelet使用cri-docker
vim /var/lib/kubelet/kubeadm-flags.env
KUBELET_KUBEADM_ARGS="--pod-infra-container-image=reg.westos.org/k8s/pause:3.7 --container-runtime=remote --container-runtime-endpoint=unix:///var/run/cri-dockerd.sock"
修改节点套接字 #需要在master节点执行
kubectl edit nodes k8s2
...
kubeadm.alpha.kubernetes.io/cri-socket: unix:///var/run/cri-dockerd.sock
重启kubelet
systemctl daemon-reload
systemctl restart kubelet 
解除节点保护 #需要在master节点执行
kubectl uncordon k8s2
其它节点依此类推
完成升级
4. 部署containerd
k8s从1.24版本开始移除了dockershim,所以我们不在使用docker，选用containerd。
k8s1、k8s2、k8s3在配置前需要重置节点
kubeadm reset
kubeadm reset --cri-socket unix:///var/run/cri-dockerd.sock
kubeadm reset --cri-socket unix:///var/run/cri-dockerd.sock
k8s1:
k8s2,3
所有节点清除iptables规则
iptables -F
iptables -F -t nat
禁用所有节点docker和cri-docker服务
systemctl disable --now docker
systemctl disable --now docker.socket
systemctl disable --now cri-docker
之前部署过docker，containerd默认已经安装
修改配置
containerd config default | tee /etc/containerd/config.toml
cd /etc/containerd/
vim config.toml
...
sandbox_image = "reg.westos.org/k8s/pause:3.7"
...
SystemdCgroup = true
修改配置文件
vim /etc/containerd/config.toml
...
[plugins."io.containerd.grpc.v1.cri".registry]config_path = "/etc/containerd/certs.d"
mkdir -p /etc/containerd/certs.d/docker.iovim /etc/containerd/certs.d/docker.io/hosts.tomlserver = "https://registry-1.docker.io"[host."https://reg.westos.org"]capabilities = ["pull", "resolve", "push"]skip_verify = true
拷贝证书
mkdir -p /etc/containerd/certs.d/reg.westos.org
cp /etc/docker/certs.d/reg.westos.org/ca.crt /etc/containerd/certs.d/reg.westos.org/
systemctl restart containerd
scp -r certs.d/ config.toml k8s2:/etc/containerd/
scp -r certs.d/ config.toml k8s3:/etc/containerd/k8s2,3
systemctl disable --now docker cri-docker docker.socket
systemctl enable --now containerd
crictl config runtime-endpoint unix:///run/containerd/containerd.sock
启动containerd
systemctl enable --now containerd
systemctl restart containerd
crictl config runtime-endpoint unix:///run/containerd/containerd.sock
crictl img
crictl pull reg.westos.org/k8s/pause:3.6
集群初始化
kubeadm init --pod-network-cidr=10.244.0.0/16 --image-repository reg.westos.org/k8s --kubernetes-version v1.24.17
kubectl apply -f kube-flannel.yml 
