运维之k8s集群搭建

用kubeadm部署K8s群集

前提摘要：        1、各节点时间同步；         2、各节点主机名称解析：dns OR hosts 3 /etc/hosts        3、各节点iptables及firewalld服务被disable

1：master(192.168.1.179) 组件:

https://mirrors.aliyun.com/kubernetes/yum/repos/kubernetes-el7-x86_64/

配置yum仓库导入docker 下载docker-ce.repowget https://mirrors.aliyun.com/docker-ce/linux/centos/docker-ce.repocd /etc/yum.repos.d/vim k8s.repo   #配置kubetnetes仓库[k8s]name=K8s Repobaseurl=https://mirrors.aliyun.com/kubernetes/yum/repos/kubernetes-el7-x86_64/gpgcheck=0gpgkey=https://mirrors.aliyun.com/kubernetes/yum/doc/yum-key.gpgenabled=1yum repolist 查看两个仓库是否能用

安装docker-ce kubelet kubeadm kubectl

yum install docker-ce kubelet kubeadm kubectl -y 如果有报错 手动下载密钥执行下面命令wget https://mirrors.aliyun.com/kubernetes/yum/doc/yum-key.gpgrpm --import yum-key.gpg

初始化docker启动docker服务

 systemctl daemon-reload #重新加载文件 systemctl start docker.service  #启动docker systemctl enable docker.service  #设置开机自启动 systemctl enable kubelet.service  docker需要自动到docker仓库去下载镜像或者通过本地仓库下载所依赖的每个镜像文件 1、通过docker自动下载依赖镜像 cd /usr/lib/systemd/system/docker.service 在 # for containers run by docker  底下加入环境变量 Environment="HTTPS_PROXY=http://www.ik8s.io:10080" 通过代理去访问需要加载的镜像文件 Environment="NO_PROXY=127.0.0.0/8,172.20.0.0/16" 2、手动下载依赖镜像用 kubeadm config images listk8s.gcr.io/kube-apiserver:v1.13.1k8s.gcr.io/kube-controller-manager:v1.13.1k8s.gcr.io/kube-scheduler:v1.13.1k8s.gcr.io/kube-proxy:v1.13.1k8s.gcr.io/pause:3.1k8s.gcr.io/etcd:3.2.24k8s.gcr.io/coredns:1.2.6下载命令docker pull docker.io/mirrorgooglecontainers/kube-apiserver:v1.13.1       下载镜像docker tag docker.io/mirrorgooglecontainers/kube-apiserver:v1.13.1 k8s.gcr.io/kube-apiserver:v1.13.1         给镜像打标签docker pull docker.io/mirrorgooglecontainers/kube-controller-manager:v1.13.1docker tag docker.io/mirrorgooglecontainers/kube-controller-manager:v1.13.1 k8s.gcr.io/kube-controller-manager:v1.13.1docker pull docker.io/mirrorgooglecontainers/kube-scheduler:v1.13.1docker tag docker.io/mirrorgooglecontainers/kube-scheduler:v1.13.1 k8s.gcr.io/kube-scheduler:v1.13.1docker pull docker.io/mirrorgooglecontainers/kube-proxy:v1.13.1docker tag docker.io/mirrorgooglecontainers/kube-proxy:v1.13.1 k8s.gcr.io/kube-proxy:v1.13.1docker pull docker.io/mirrorgooglecontainers/pause:3.1docker tag docker.io/mirrorgooglecontainers/pause:3.1 k8s.gcr.io/pause:3.1docker pull docker.io/mirrorgooglecontainers/etcd:3.2.24docker tag docker.io/mirrorgooglecontainers/etcd:3.2.24 k8s.gcr.io/etcd:3.2.24docker pull docker.io/coredns/coredns:1.2.6docker tag docker.io/coredns/coredns:1.2.6 k8s.gcr.io/coredns:1.2.6----------------------------------------------------------------------------- 

初始化kubelet

编辑kubelet的配置文件/etc/sysconfig/kubelet，设置其忽略Swap启用的状态错误，内容如下： KUBELET_EXTRA_ARGS="--fail-swap-on=false" 设置kubelet开机自启动 systemctl enable kubelet.service 初始化 kubeadm init --kubernetes-version=v1.13.1 --pod-network-cidr=192.168.0.0/6 --service.cidr=192.168.1.0/12 --ignore-preflight-errors=Swap --kubernetes-version=v1.11.1     #被初始化的k8s版本 --pod-network-cidr=10.244.0.0/16 #pod使用的网络 --service.cidr=10.96.0.0/12      #service使用的网络地址 --ignore-preflight-errors=Swap   #双重否定部属用用交换配置 如果在初始化过程中有error报错那么就查看是什么错误一般会报错镜像tag标签错误找不到该镜像  可以把镜像重新打标签

当如果看到下列内容则表示初始化成功

Your Kubernetes master has initialized successfully!To start using your cluster, you need to run the following as a regular user:  mkdir ~/.kube          #创建.kube文件一般会提示用普通用户创建要有sudo权限这边我直接用root用户创建  sudo cp /etc/kubernetes/admin.conf ~/.kube/ #复制admin.conf到./kube文件中  sudo chown $(id -u):$(id -g) $HOME/.kube/config #把./kube的属组属主改为创建用户的属组属主You should now deploy a pod network to the cluster.Run "kubectl apply -f [podnetwork].yaml" with one of the options listed at:  https://kubernetes.io/docs/concepts/cluster-administration/addons/You can now join any number of machines by running the following on each nodeas root:#用下列命令把node加入集群中来（建议把下列命令保存起来）  kubeadm join 192.168.1.179:6443 --token 5m7gg1.czd5td6itn9g2fhz --discovery-token-ca-cert-hash sha256:50c64cac88defae6beecf7bdde9b212094d7cc937b709b94f0baaeaaa4246e7e用root用户创建：mkdir ~/.kubecp /etc/kubernetes/admin.conf ~/.kube/因为应root用户创建的就不用重新指定属组属主kubectl get cs   #查看组件健康状态kubectl get nodes #查看个节点信息NAME     STATUS   ROLES    AGE   VERSIONmaster   NotReady    master   19h   v1.13.1   #这边显示未开启状态因为这边少一个flannel插件安装flannel插件https://github.com/coreos/flannel #flannel地址和帮助信息图片#如果你要受用安装就执行下面的命令就会自动执行安装kubectl apply -f https://raw.githubusercontent.com/coreos/flannel/master/Documentation/kube-flannel.yml#部署完之后再查看节点信息NAME     STATUS   ROLES    AGE   VERSIONmaster   Ready    master   19h   v1.13.1#查看组件状态信息（当前master运行的所有pod组件）kubectl get pods -n kube-systemNAME                             READY   STATUS    RESTARTS   AGEcoredns-86c58d9df4-4frfz         1/1     Running   0          19hcoredns-86c58d9df4-hlr72         1/1     Running   0          19hetcd-master                      1/1     Running   0          19hkube-apiserver-master            1/1     Running   0          19hkube-controller-manager-master   1/1     Running   0          19hkube-flannel-ds-amd64-4c7jx      1/1     Running   0          19hkube-flannel-ds-amd64-89m8l      1/1     Running   0          17hkube-flannel-ds-amd64-rmxj9      1/1     Running   0          19hkube-proxy-8pnqs                 1/1     Running   0          17hkube-proxy-b4hlj                 1/1     Running   0          19hkube-proxy-fzp2m                 1/1     Running   0          19hkube-scheduler-master            1/1     Running   0          19h#默认如果不指定名名称空间则名称空间kube-systemkubectl get nsNAME          STATUS   AGEdefault       Active   19hkube-public   Active   19hkube-system   Active   19h#主节点初始化完成

node01 安装初始化

如果用第一种方法让docker自动下载依赖镜像则复制master中docker.service文件到node01中
把master中的k8s.repo和下载的docker-ce的yum仓库配置起来如果在yum过程中有报错则按照master中的方法执行

yum install docker-ce kubelet kubeadm -y1、开启docker服务2、把docker 、 kubelet设置开机自启动 systemctl start docker.service  #启动docker systemctl enable docker.service  #设置开机自启动 systemctl enable kubelet.service 手动下载kube-proxy 、 pause 镜像并打标签通master中的下载镜像配置把node01节点加入集群中去用刚刚保存的kubeadm job命令kubeadm join 192.168.1.179:6443 --token 5m7gg1.czd5td6itn9g2fhz --discovery-token-ca-cert-hash sha256:50c64cac88defae6beecf7bdde9b212094d7cc937b709b94f0baaeaaa4246e7e --ignore-preflight-errors=Swap #后面加入这句话

再主节点上kubectl get nodes 查看系欸但有没有加入进来

kubectl get nodesNAME     STATUS   ROLES    AGE   VERSIONmaster   Ready    master   20h   v1.13.1node01   Ready       19h   v1.13.1 #已经加入进来了# node02跟node01做同样步骤加入集群