一. 为什么是k8s v1.16.0?

最新版的v1.16.2试过了,一直无法安装完成,安装到kubeadm init那一步执行后,报了很多错,如:node xxx not found等。centos7都重装了几次,还是无法解决。用了一天都没安装完,差点放弃。

后来在网上搜到的安装教程基本都是v1.16.0的,我不太相信是v1.16.2的坑所以先前没打算降级到v1.16.0。没办法了就试着安装v1.16.0版本,竟然成功了。记录在此,避免后来者踩坑。

本篇文章,安装大步骤如下:

  • 安装docker-ce 18.09.9(所有机器)

  • 设置k8s环境前置条件(所有机器)

  • 安装k8s v1.16.0 master管理节点

  • 安装k8s v1.16.0 node工作节点

  • 安装flannel(master)

这里有重要的一步,请记住自己master和node之间通信的ip,如我的master的ip为192.168.99.104,node的ip为:192.168.99.105. 请确保使用这两个ip在master和node上能互相ping通,这个master的ip 192.168.99.104接下来配置k8s的时候需要用到。

我的环境:

  • 操作系统:win10

  • 虚拟机:virtual box

  • linux发行版:CentOS7

  • linux内核(使用uname -r查看):3.10.0-957.el7.x86_64

  • master和node节点通信的ip(master):192.168.99.104

二. 安装docker-ce 18.09.9(所有机器)

所有安装k8s的机器都需要安装docker,命令如下:

# 安装docker所需的工具
yum install -y yum-utils device-mapper-persistent-data lvm2
# 配置阿里云的docker源yum-config-manager --add-repo http://mirrors.aliyun.com/docker-ce/linux/centos/docker-ce.repo
# 指定安装这个版本的docker-ce
yum install -y docker-ce-18.09.9-3.el7
# 启动docker
systemctl enable docker && systemctl start docker

三. 设置k8s环境准备条件(所有机器)

安装k8s的机器需要2个CPU和2g内存以上,这个简单,在虚拟机里面配置一下就可以了。然后执行以下脚本做一些准备操作。所有安装k8s的机器都需要这一步操作。

# 关闭防火墙
systemctl disable firewalldsystemctl stop firewalld# 关闭selinux# 临时禁用selinuxsetenforce 0
# 永久关闭 修改/etc/sysconfig/selinux文件设置sed -i 's/SELINUX=permissive/SELINUX=disabled/' /etc/sysconfig/selinux
sed -i "s/SELINUX=enforcing/SELINUX=disabled/g" /etc/selinux/config
# 禁用交换分区swapoff -a# 永久禁用,打开/etc/fstab注释掉swap那一行。sed -i 's/.*swap.*/#&/' /etc/fstab
# 修改内核参数cat <<EOF >  /etc/sysctl.d/k8s.confnet.bridge.bridge-nf-call-ip6tables = 1
net.bridge.bridge-nf-call-iptables = 1
EOFsysctl --system

四. 安装k8s v1.16.0 master管理节点

如果还没安装docker,请参照本文步骤二安装docker-ce 18.09.9(所有机器)安装。

如果没设置k8s环境准备条件,请参照本文步骤三设置k8s环境准备条件(所有机器)执行。

以上两个步骤检查完毕之后,继续以下步骤。

1. 安装kubeadm、kubelet、kubectl

由于官方k8s源在google,国内无法访问,这里使用阿里云yum源

# 执行配置k8s阿里云源
cat <<EOF > /etc/yum.repos.d/kubernetes.repo
[kubernetes]
name=Kubernetes
baseurl=https://mirrors.aliyun.com/kubernetes/yum/repos/kubernetes-el7-x86_64/
enabled=1
gpgcheck=1
repo_gpgcheck=1
gpgkey=https://mirrors.aliyun.com/kubernetes/yum/doc/yum-key.gpg https://mirrors.aliyun.com/kubernetes/yum/doc/rpm-package-key.gpg
EOF
# 安装kubeadm、kubectl、kubelet
yum install -y kubectl-1.16.0-0 kubeadm-1.16.0-0 kubelet-1.16.0-0
# 启动kubelet服务
systemctl enable kubelet && systemctl start kubelet

2. 初始化k8s

以下这个命令开始安装k8s需要用到的docker镜像,因为无法访问到国外网站,所以这条命令使用的是国内的阿里云的源(registry.aliyuncs.com/google_containers)。

另一个非常重要的是:这里的--apiserver-advertise-address使用的是master和node间能互相ping通的ip,我这里是192.168.99.104,刚开始在这里被坑了一个晚上,你请自己修改下ip执行。

这条命令执行时会卡在[preflight] You can also perform this action in beforehand using ''kubeadm config images pull,大概需要2分钟,请耐心等待。

# 下载管理节点中用到的6个docker镜像,你可以使用docker images查看到
# 这里需要大概两分钟等待,会卡在[preflight] You can also perform this action in beforehand using ''kubeadm config images pull
kubeadm init --image-repository registry.aliyuncs.com/google_containers --kubernetes-version v1.16.0 --apiserver-advertise-address 192.168.99.104 --pod-network-cidr=10.244.0.0/16 --token-ttl 0

上面安装完后,会提示你输入如下命令,复制粘贴过来,执行即可。

# 上面安装完成后,k8s会提示你输入如下命令,执行
mkdir -p $HOME/.kube
sudo cp -i /etc/kubernetes/admin.conf $HOME/.kube/config
sudo chown $(id -u):$(id -g) $HOME/.kube/config

3. 记住node加入集群的命令

上面kubeadm init执行成功后会返回给你node节点加入集群的命令,等会要在node节点上执行,需要保存下来,如果忘记了,可以使用如下命令获取。

kubeadm token create --print-join-command 

以上,安装master节点完毕。可以使用kubectl get nodes查看一下,此时master处于NotReady状态,暂时不用管。

五. 安装k8s v1.16.0 node工作节点

如果还没安装docker,请参照本文步骤二安装docker-ce 18.09.9(所有机器)安装。

如果没设置k8s环境准备条件,请参照本文步骤三设置k8s环境准备条件(所有机器)执行。

以上两个步骤检查完毕之后,继续以下步骤。

1. 安装kubeadm、kubelet

# 执行配置k8s阿里云源
cat <<EOF > /etc/yum.repos.d/kubernetes.repo
[kubernetes]
name=Kubernetes
baseurl=https://mirrors.aliyun.com/kubernetes/yum/repos/kubernetes-el7-x86_64/
enabled=1
gpgcheck=1
repo_gpgcheck=1
gpgkey=https://mirrors.aliyun.com/kubernetes/yum/doc/yum-key.gpg https://mirrors.aliyun.com/kubernetes/yum/doc/rpm-package-key.gpg
EOF
# 安装kubeadm、kubectl、kubelet
yum install -y  kubeadm-1.16.0-0 kubelet-1.16.0-0
# 启动kubelet服务
systemctl enable kubelet && systemctl start kubelet

2. 加入集群

这里加入集群的命令每个人都不一样,可以登录master节点,使用kubeadm token create --print-join-command 来获取。获取后执行如下。

# 加入集群,如果这里不知道加入集群的命令,可以登录master节点,使用kubeadm token create --print-join-command 来获取 
kubeadm join 192.168.99.104:6443 --token ncfrid.7ap0xiseuf97gikl \
    --discovery-token-ca-cert-hash sha256:47783e9851a1a517647f1986225f104e81dbfd8fb256ae55ef6d68ce9334c6a2

加入成功后,可以在master节点上使用kubectl get nodes命令查看到加入的节点。

六. 安装flannel(master机器)

以上步骤安装完后,机器搭建起来了,但状态还是NotReady状态,如下图,master机器需要安装flanneld。

1. 下载官方fannel配置文件

使用wget命令,地址为:(https://raw.githubusercontent.com/coreos/flannel/master/Documentation/kube-flannel.yml),这个地址国内访问不了,所以我把内容复制下来,为了避免前面文章过长,我把它粘贴到文章末尾第八个步骤附录了。

这个yml配置文件中配置了一个国内无法访问的地址(quay.io),我已经将其改为国内可以访问的地址(quay-mirror.qiniu.com)。我们新建一个kube-flannel.yml文件,复制粘贴该内容即可。

搜索Java知音公众号,回复“后端面试”,送你一份Java面试题宝典.pdf

2. 安装fannel

kubectl apply -f kube-flannel.yml

七. 大功告成

至此,k8s集群搭建完成,如下图节点已为Ready状态,大功告成,完结撒花。

八. 附录

这是kube-flannel.yml文件的内容,已经将无法访问的地址(quay.io)全部改为国内可以访问的地址(quay-mirror.qiniu.com)。我们新建一个kube-flannel.yml文件,复制粘贴该内容即可。

---
apiVersion: policy/v1beta1kind: PodSecurityPolicymetadata:  name: psp.flannel.unprivileged  annotations:    seccomp.security.alpha.kubernetes.io/allowedProfileNames: docker/default
    seccomp.security.alpha.kubernetes.io/defaultProfileName: docker/default
    apparmor.security.beta.kubernetes.io/allowedProfileNames: runtime/default
    apparmor.security.beta.kubernetes.io/defaultProfileName: runtime/defaultspec:  privileged: false
  volumes:    - configMap    - secret    - emptyDir    - hostPath  allowedHostPaths:    - pathPrefix: "/etc/cni/net.d"
    - pathPrefix: "/etc/kube-flannel"
    - pathPrefix: "/run/flannel"
  readOnlyRootFilesystem: false
  # Users and groups  runAsUser:    rule: RunAsAny  supplementalGroups:    rule: RunAsAny  fsGroup:    rule: RunAsAny  # Privilege Escalation  allowPrivilegeEscalation: false
  defaultAllowPrivilegeEscalation: false
  # Capabilities  allowedCapabilities: ['NET_ADMIN']
  defaultAddCapabilities: []  requiredDropCapabilities: []  # Host namespaces  hostPID: false
  hostIPC: false
  hostNetwork: true
  hostPorts:  - min: 0
    max: 65535
  # SELinux  seLinux:    # SELinux is unused in CaaSP
    rule: 'RunAsAny'
---kind: ClusterRoleapiVersion: rbac.authorization.k8s.io/v1beta1metadata:  name: flannelrules:  - apiGroups: ['extensions']
    resources: ['podsecuritypolicies']
    verbs: ['use']
    resourceNames: ['psp.flannel.unprivileged']
  - apiGroups:      - ""
    resources:      - pods    verbs:      - get
  - apiGroups:      - ""
    resources:      - nodes    verbs:      - list      - watch  - apiGroups:      - ""
    resources:      - nodes/status    verbs:      - patch---kind: ClusterRoleBindingapiVersion: rbac.authorization.k8s.io/v1beta1metadata:  name: flannelroleRef:  apiGroup: rbac.authorization.k8s.io  kind: ClusterRole  name: flannelsubjects:- kind: ServiceAccount  name: flannel  namespace: kube-system
---apiVersion: v1kind: ServiceAccountmetadata:  name: flannel  namespace: kube-system
---kind: ConfigMapapiVersion: v1metadata:  name: kube-flannel-cfg  namespace: kube-system
  labels:    tier: node    app: flanneldata:  cni-conf.json: |    {      "name""cbr0",
      "cniVersion""0.3.1",
      "plugins": [
        {          "type""flannel",
          "delegate": {
            "hairpinMode"true,
            "isDefaultGateway"true
          }        },        {          "type""portmap",
          "capabilities": {
            "portMappings"true
          }        }      ]    }  net-conf.json: |    {      "Network""10.244.0.0/16",
      "Backend": {
        "Type""vxlan"
      }    }---apiVersion: apps/v1kind: DaemonSetmetadata:  name: kube-flannel-ds-amd64  namespace: kube-system
  labels:    tier: node    app: flannelspec:  selector:    matchLabels:      app: flannel  template:    metadata:      labels:        tier: node        app: flannel    spec:      affinity:        nodeAffinity:          requiredDuringSchedulingIgnoredDuringExecution:            nodeSelectorTerms:              - matchExpressions:                  - key: beta.kubernetes.io/os                    operator: In                    values:                      - linux                  - key: beta.kubernetes.io/arch                    operator: In                    values:                      - amd64      hostNetwork: true
      tolerations:      - operator: Exists        effect: NoSchedule      serviceAccountName: flannel      initContainers:      - name: install-cni        image: quay-mirror.qiniu.com/coreos/flannel:v0.11.0-amd64
        command:        - cp        args:        - -f        - /etc/kube-flannel/cni-conf.json
        - /etc/cni/net.d/10-flannel.conflist
        volumeMounts:        - name: cni          mountPath: /etc/cni/net.d
        - name: flannel-cfg          mountPath: /etc/kube-flannel/
      containers:      - name: kube-flannel        image: quay-mirror.qiniu.com/coreos/flannel:v0.11.0-amd64
        command:        - /opt/bin/flanneld
        args:        - --ip-masq        - --kube-subnet-mgr        resources:          requests:            cpu: "100m"
            memory: "50Mi"
          limits:            cpu: "100m"
            memory: "50Mi"
        securityContext:          privileged: false
          capabilities:            add: ["NET_ADMIN"]
        env:        - name: POD_NAME          valueFrom:            fieldRef:              fieldPath: metadata.name        - name: POD_NAMESPACE          valueFrom:            fieldRef:              fieldPath: metadata.namespace        volumeMounts:        - name: run          mountPath: /run/flannel
        - name: flannel-cfg          mountPath: /etc/kube-flannel/
      volumes:        - name: run          hostPath:            path: /run/flannel
        - name: cni          hostPath:            path: /etc/cni/net.d
        - name: flannel-cfg          configMap:            name: kube-flannel-cfg---apiVersion: apps/v1kind: DaemonSetmetadata:  name: kube-flannel-ds-arm64  namespace: kube-system
  labels:    tier: node    app: flannelspec:  selector:    matchLabels:      app: flannel  template:    metadata:      labels:        tier: node        app: flannel    spec:      affinity:        nodeAffinity:          requiredDuringSchedulingIgnoredDuringExecution:            nodeSelectorTerms:              - matchExpressions:                  - key: beta.kubernetes.io/os                    operator: In                    values:                      - linux                  - key: beta.kubernetes.io/arch                    operator: In                    values:                      - arm64      hostNetwork: true
      tolerations:      - operator: Exists        effect: NoSchedule      serviceAccountName: flannel      initContainers:      - name: install-cni        image: quay-mirror.qiniu.com/coreos/flannel:v0.11.0-arm64
        command:        - cp        args:        - -f        - /etc/kube-flannel/cni-conf.json
        - /etc/cni/net.d/10-flannel.conflist
        volumeMounts:        - name: cni          mountPath: /etc/cni/net.d
        - name: flannel-cfg          mountPath: /etc/kube-flannel/
      containers:      - name: kube-flannel        image: quay-mirror.qiniu.com/coreos/flannel:v0.11.0-arm64
        command:        - /opt/bin/flanneld
        args:        - --ip-masq        - --kube-subnet-mgr        resources:          requests:            cpu: "100m"
            memory: "50Mi"
          limits:            cpu: "100m"
            memory: "50Mi"
        securityContext:          privileged: false
          capabilities:             add: ["NET_ADMIN"]
        env:        - name: POD_NAME          valueFrom:            fieldRef:              fieldPath: metadata.name        - name: POD_NAMESPACE          valueFrom:            fieldRef:              fieldPath: metadata.namespace        volumeMounts:        - name: run          mountPath: /run/flannel
        - name: flannel-cfg          mountPath: /etc/kube-flannel/
      volumes:        - name: run          hostPath:            path: /run/flannel
        - name: cni          hostPath:            path: /etc/cni/net.d
        - name: flannel-cfg          configMap:            name: kube-flannel-cfg---apiVersion: apps/v1kind: DaemonSetmetadata:  name: kube-flannel-ds-arm  namespace: kube-system
  labels:    tier: node    app: flannelspec:  selector:    matchLabels:      app: flannel  template:    metadata:      labels:        tier: node        app: flannel    spec:      affinity:        nodeAffinity:          requiredDuringSchedulingIgnoredDuringExecution:            nodeSelectorTerms:              - matchExpressions:                  - key: beta.kubernetes.io/os                    operator: In                    values:                      - linux                  - key: beta.kubernetes.io/arch                    operator: In                    values:                      - arm      hostNetwork: true
      tolerations:      - operator: Exists        effect: NoSchedule      serviceAccountName: flannel      initContainers:      - name: install-cni        image: quay-mirror.qiniu.com/coreos/flannel:v0.11.0-arm
        command:        - cp        args:        - -f        - /etc/kube-flannel/cni-conf.json
        - /etc/cni/net.d/10-flannel.conflist
        volumeMounts:        - name: cni          mountPath: /etc/cni/net.d
        - name: flannel-cfg          mountPath: /etc/kube-flannel/
      containers:      - name: kube-flannel        image: quay-mirror.qiniu.com/coreos/flannel:v0.11.0-arm
        command:        - /opt/bin/flanneld
        args:        - --ip-masq        - --kube-subnet-mgr        resources:          requests:            cpu: "100m"
            memory: "50Mi"
          limits:            cpu: "100m"
            memory: "50Mi"
        securityContext:          privileged: false
          capabilities:             add: ["NET_ADMIN"]
        env:        - name: POD_NAME          valueFrom:            fieldRef:              fieldPath: metadata.name        - name: POD_NAMESPACE          valueFrom:            fieldRef:              fieldPath: metadata.namespace        volumeMounts:        - name: run          mountPath: /run/flannel
        - name: flannel-cfg          mountPath: /etc/kube-flannel/
      volumes:        - name: run          hostPath:            path: /run/flannel
        - name: cni          hostPath:            path: /etc/cni/net.d
        - name: flannel-cfg          configMap:            name: kube-flannel-cfg---apiVersion: apps/v1kind: DaemonSetmetadata:  name: kube-flannel-ds-ppc64le  namespace: kube-system
  labels:    tier: node    app: flannelspec:  selector:    matchLabels:      app: flannel  template:    metadata:      labels:        tier: node        app: flannel    spec:      affinity:        nodeAffinity:          requiredDuringSchedulingIgnoredDuringExecution:            nodeSelectorTerms:              - matchExpressions:                  - key: beta.kubernetes.io/os                    operator: In                    values:                      - linux                  - key: beta.kubernetes.io/arch                    operator: In                    values:                      - ppc64le      hostNetwork: true
      tolerations:      - operator: Exists        effect: NoSchedule      serviceAccountName: flannel      initContainers:      - name: install-cni        image: quay-mirror.qiniu.com/coreos/flannel:v0.11.0-ppc64le
        command:        - cp        args:        - -f        - /etc/kube-flannel/cni-conf.json
        - /etc/cni/net.d/10-flannel.conflist
        volumeMounts:        - name: cni          mountPath: /etc/cni/net.d
        - name: flannel-cfg          mountPath: /etc/kube-flannel/
      containers:      - name: kube-flannel        image: quay-mirror.qiniu.com/coreos/flannel:v0.11.0-ppc64le
        command:        - /opt/bin/flanneld
        args:        - --ip-masq        - --kube-subnet-mgr        resources:          requests:            cpu: "100m"
            memory: "50Mi"
          limits:            cpu: "100m"
            memory: "50Mi"
        securityContext:          privileged: false
          capabilities:             add: ["NET_ADMIN"]
        env:        - name: POD_NAME          valueFrom:            fieldRef:              fieldPath: metadata.name        - name: POD_NAMESPACE          valueFrom:            fieldRef:              fieldPath: metadata.namespace        volumeMounts:        - name: run          mountPath: /run/flannel
        - name: flannel-cfg          mountPath: /etc/kube-flannel/
      volumes:        - name: run          hostPath:            path: /run/flannel
        - name: cni          hostPath:            path: /etc/cni/net.d
        - name: flannel-cfg          configMap:            name: kube-flannel-cfg---apiVersion: apps/v1kind: DaemonSetmetadata:  name: kube-flannel-ds-s390x  namespace: kube-system
  labels:    tier: node    app: flannelspec:  selector:    matchLabels:      app: flannel  template:    metadata:      labels:        tier: node        app: flannel    spec:      affinity:        nodeAffinity:          requiredDuringSchedulingIgnoredDuringExecution:            nodeSelectorTerms:              - matchExpressions:                  - key: beta.kubernetes.io/os                    operator: In                    values:                      - linux                  - key: beta.kubernetes.io/arch                    operator: In                    values:                      - s390x      hostNetwork: true
      tolerations:      - operator: Exists        effect: NoSchedule      serviceAccountName: flannel      initContainers:      - name: install-cni        image: quay-mirror.qiniu.com/coreos/flannel:v0.11.0-s390x
        command:        - cp        args:        - -f        - /etc/kube-flannel/cni-conf.json
        - /etc/cni/net.d/10-flannel.conflist
        volumeMounts:        - name: cni          mountPath: /etc/cni/net.d
        - name: flannel-cfg          mountPath: /etc/kube-flannel/
      containers:      - name: kube-flannel        image: quay-mirror.qiniu.com/coreos/flannel:v0.11.0-s390x
        command:        - /opt/bin/flanneld
        args:        - --ip-masq        - --kube-subnet-mgr        resources:          requests:            cpu: "100m"
            memory: "50Mi"
          limits:            cpu: "100m"
            memory: "50Mi"
        securityContext:          privileged: false
          capabilities:             add: ["NET_ADMIN"]
        env:        - name: POD_NAME          valueFrom:            fieldRef:              fieldPath: metadata.name        - name: POD_NAMESPACE          valueFrom:            fieldRef:              fieldPath: metadata.namespace        volumeMounts:        - name: run          mountPath: /run/flannel
        - name: flannel-cfg          mountPath: /etc/kube-flannel/
      volumes:        - name: run          hostPath:            path: /run/flannel
        - name: cni          hostPath:            path: /etc/cni/net.d
        - name: flannel-cfg          configMap:            name: kube-flannel-cfg