安装
环境预准备
-
安装 docker：
curl -sSL https://get.daocloud.io/docker | sh systemctl enable docker systemctl start docker -
更改 Docker 源以及保持 Docker Cgroup Driver 和 k8s 一致:
cat <<EOF > /etc/docker/deamon.json { "registry-mirrors": ["https://registry.cn-hangzhou.aliyuncs.com"], "exec-opts": ["native.cgroupdriver=systemd"] } EOF systemctl enable docker systemctl start docker -
关闭 selinux:
# setenforce 0 # vim /etc/sysconfig/selinux SELINUX=disabled -
关闭交换分区:
swapoff -a vim /etc/fstab # 注释掉 swap 行 -
iptables 配置：
cat <<EOF | sudo tee /etc/modules-load.d/k8s.conf br_netfilter EOF cat <<EOF | sudo tee /etc/sysctl.d/k8s.conf net.bridge.bridge-nf-call-ip6tables = 1 net.bridge.bridge-nf-call-iptables = 1 EOF sudo sysctl --system -
关闭防火墙：
systemctl disable firewalld systemctl stop firewalld
开始安装
安装 kublet,kubeadm,kubctl，版本都是1.18.3。
-
yum 设置源：
cat <<EOF > /etc/yum.repos.d/kubernetes.repo [kubernetes] name=Kubernetes Respository baseurl=http://mirrors.aliyun.com/kubernetes/yum/repos/kubernetes-el7-x86_64/ enabled=1 gpgcheck=0 EOF -
安装 kublet,kubeadm,kubctl:
yum install -y kubelet-1.18.3 kubectl-1.18.3 kubeadm-1.18.3 --disableexcludes=kubernetes systemctl enable kubelet && systemctl start kubelet -
首先在Master上安装，安装前更改配置文件：
# kubeadm config print init-defaults > init.config.yaml # vim init.config.yaml ... localAPIEndpoint: advertiseAddress: 1.2.3.4 ... # 1.2.3.4 修改为本机物理网卡ip地址，只有Master安装时需要改 ... imageRepository: k8s.gcr.io ... # k8s.gcr.io 改为 registry.cn-hangzhou.aliyuncs.com/google_containers -
查看镜像地址发现还是指向 k8s.grc.io:
kubeadm config images list -
可以使用之前更改过的配置文件提前拉取 kubeadm 所需镜像:
kubeadm config images pull --config=init.config.yaml -
开始安装Master:
kubeadm init --config=init.config.yaml -
完成后按照提示执行下面命令：
mkdir -p $HOME/.kube sudo cp -i /etc/kubernetes/admin.conf $HOME/.kube/config sudo chown $(id -u):$(id -g) $HOME/.kube/config -
还有一个Node加入的token命令，这里保存下来，类似：
kubeadm join 192.168.10.8:6443 --token y3r86i.anh1x5pxt2mc680y \ --discovery-token-ca-cert-hash sha256:9762cd94ffe6c35089d647e8b7a817226b57f21d19cafc977f41140cfaec19b9注意：如果时间过久（Node加入时间与这个token的创建时间间隔），token会失效（默认24小时），这时可以
kubeadm token create --ttl 0创建永久token或者kubeadm token create创建默认24小时有效时间的token，替换上面的token之后再join即可。 -
开始安装Node，执行上述第1,2,3,4,5步，之后执行第8步保存下来的命令。
-
为了能够在Node上直接使用kubectl（不显示指定conf文件）：
cp /etc/kubernetes/kubelet.conf /etc/kubernetes/admin.conf -
此时在Master上输入
kubectl get nodes，发现 STATUS 全部是 NotReady，这是因为没有安装 cni 网络插件:[root@localhost ~]# kubectl get nodes NAME STATUS ROLES AGE VERSION centos2 NotReady <none> 2m53s v1.18.3 centos3 NotReady <none> 117s v1.18.3 localhost.localdomain NotReady master 5h52m v1.18.3
安装网络插件
在Master上安装 calico。
-
下载 yaml 文件：
wget https://docs.projectcalico.org/manifests/calico.yaml -
vim 查找
docker.io发现下面四个镜像：calico/cni:v3.21.0 calico/pod2daemon-flexvol:v3.21.0 calico/node:v3.21.0 calico/kube-controllers:v3.21.0 -
docker 先用之前替换好的国内源把镜像拉下来：
docker pull calico/cni:v3.21.0 && \ docker pull calico/pod2daemon-flexvol:v3.21.0 && \ docker pull calico/node:v3.21.0 && \ docker pull calico/kube-controllers:v3.21.0 -
部署 calico：
kubectl apply -f calico.yaml -
此时查看 pods，会看到 calico 正在初始化，等待完成后，STATUS 会全部变成 Running：
[root@localhost ~]# kubectl get pods --all-namespaces NAMESPACE NAME READY STATUS RESTARTS AGE kube-system calico-kube-controllers-858fbfbc9-mpnjt 1/1 Running 0 96s kube-system calico-node-p265c 1/1 Running 0 97s kube-system calico-node-xrsgm 0/1 PodInitializing 0 97s kube-system calico-node-z946n 0/1 PodInitializing 0 97s kube-system coredns-546565776c-cxnnf 1/1 Running 0 7h3m kube-system coredns-546565776c-f299w 1/1 Running 0 7h3m kube-system etcd-localhost.localdomain 1/1 Running 0 7h3m kube-system kube-apiserver-localhost.localdomain 1/1 Running 0 7h3m kube-system kube-controller-manager-localhost.localdomain 1/1 Running 0 7h3m kube-system kube-proxy-777wz 1/1 Running 0 74m kube-system kube-proxy-mj6r8 1/1 Running 0 7h3m kube-system kube-proxy-tx56l 1/1 Running 0 73m kube-system kube-scheduler-localhost.localdomain 1/1 Running 0 7h3m -
查看 nodes，STATUS 已经全部 Ready:
[root@localhost ~]# kubectl get nodes NAME STATUS ROLES AGE VERSION centos2 Ready <none> 76m v1.18.3 centos3 Ready <none> 75m v1.18.3 localhost.localdomain Ready master 7h6m v1.18.3
错误记录
-
安装完 Master 后:
[root@localhost ~]# kubectl get nodes Unable to connect to the server: x509: certificate signed by unknown authority (possibly because of "crypto/rsa: verification error" while trying to verify candidate authority certificate "kubernetes")解决：
[root@localhost ~]# export KUBECONFIG=/etc/kubernetes/kubelet.conf [root@localhost ~]# kubectl get nodes NAME STATUS ROLES AGE VERSION localhost.localdomain Ready master 2m20s v1.18.3 -
安装 calico 时出错(RBAC)：
[root@localhost ~]# kubectl apply -f calico.yaml Error from server (Forbidden): error when retrieving current configuration of: Resource: "/v1, Resource=configmaps", GroupVersionKind: "/v1, Kind=ConfigMap" Name: "calico-config", Namespace: "kube-system"............................. -
在 node 上执行
kubectl get pods出现:[root@centos3 ~]# kubectl get pods The connection to the server localhost:8080 was refused - did you specify the right host or port?解决：
mkdir -p /root/.kube && cp /etc/kubernetes/kubelet.conf /root/.kube/config