Kubernetes

ansible-playbook部署K8S高可用集群

本文主要是介绍ansible-playbook部署K8S高可用集群,对大家解决编程问题具有一定的参考价值,需要的程序猿们随着小编来一起学习吧!

通过ansible-playbook,以Kubeadm方式部署K8S高可用集群(多主多从)。

kubernetes安装目录:  /etc/kubernetes/

KubeConfig:	~/.kube/config

Version:	 v1.14.0

主机说明:

系统ip角色cpu内存hostname
CentOS 7.7192.168.30.128master>=2>=2Gmaster1
CentOS 7.7192.168.30.129master>=2>=2Gmaster2
CentOS 7.7192.168.30.130node>=2>=2Gnode1
CentOS 7.7192.168.30.131node>=2>=2Gnode2
CentOS 7.7192.168.30.132node>=2>=2Gnode3

准备

  • 将所有部署k8s集群的主机分组:
# vim /etc/ansible/hosts[master]192.168.30.128 hostname=master1[add_master]192.168.30.129 hostname=master2[add_node]192.168.30.130 hostname=node1
192.168.30.131 hostname=node2
192.168.30.132 hostname=node3

  • 创建管理目录:
# mkdir -p kubeadm_k8s/roles/{docker_install,init_install,master_install,node_install,addons_install}/{files,handlers,meta,tasks,templates,vars}# cd kubeadm_k8s/

说明:

files:存放需要同步到异地服务器的源码文件及配置文件; 
handlers:当资源发生变化时需要进行的操作,若没有此目录可以不建或为空; 
meta:存放说明信息、说明角色依赖等信息,可留空; 
tasks:K8S 安装过程中需要进行执行的任务; 
templates:用于执行 K8S 安装的模板文件,一般为脚本; 
vars:本次安装定义的变量

# tree ..├── k8s.yml
└── roles
    ├── addons_install
    │   ├── files
    │   ├── handlers
    │   ├── meta
    │   ├── tasks
    │   │   ├── calico.yml
    │   │   ├── dashboard.yml
    │   │   └── main.yml
    │   ├── templates
    │   │   ├── calico-rbac-kdd.yaml
    │   │   ├── calico.yaml
    │   │   └── dashboard-all.yaml
    │   └── vars
    │       └── main.yml
    ├── docker_install
    │   ├── files
    │   ├── handlers
    │   ├── meta
    │   ├── tasks
    │   │   ├── install.yml
    │   │   ├── main.yml
    │   │   └── prepare.yml
    │   ├── templates
    │   │   ├── daemon.json
    │   │   ├── kubernetes.conf
    │   │   └── kubernetes.repo
    │   └── vars
    │       └── main.yml
    ├── init_install
    │   ├── files
    │   ├── handlers
    │   ├── meta
    │   ├── tasks
    │   │   ├── install.yml
    │   │   └── main.yml
    │   ├── templates
    │   │   ├── check-apiserver.sh
    │   │   ├── keepalived-master.conf
    │   │   └── kubeadm-config.yaml
    │   └── vars
    │       └── main.yml
    ├── master_install
    │   ├── files
    │   ├── handlers
    │   ├── meta
    │   ├── tasks
    │   │   ├── install.yml
    │   │   └── main.yml
    │   ├── templates
    │   │   ├── check-apiserver.sh
    │   │   └── keepalived-backup.conf
    │   └── vars
    │       └── main.yml
    └── node_install
        ├── files
        ├── handlers
        ├── meta
        ├── tasks
        │   ├── install.yml
        │   └── main.yml
        ├── templates
        └── vars
            └── main.yml

36 directories, 29 files

  • 创建安装入口文件,用来调用roles:
# vim k8s.yml---
- hosts: all
  remote_user: root
  gather_facts: True

  roles:
    - docker_install

- hosts: master
  remote_user: root
  gather_facts: True

  roles:
    - init_install

- hosts: add_master
  remote_user: root
  gather_facts: True

  roles:
    - master_install
    
- hosts: add_node
  remote_user: root
  gather_facts: True

  roles:
    - node_install
    
- hosts: master
  remote_user: root
  gather_facts: True

  roles:
    - addons_install


docker部分

  • 创建docker入口文件,用来调用docker_install:
# vim docker.yml#用于批量安装Docker- hosts: all
  remote_user: root
  gather_facts: True

  roles:
    - docker_install

  • 创建变量:
# vim roles/docker_install/vars/main.yml#定义docker安装中的变量SOURCE_DIR: /software
VERSION: 1.14.0-0

  • 创建模板文件:

docker配置daemon.json

# vim roles/docker_install/templates/daemon.json{  
    "registry-mirrors": ["http://f1361db2.m.daocloud.io"],    "exec-opts":["native.cgroupdriver=systemd"]}

系统环境kubernetes.conf

# vim roles/docker_install/templates/kubernetes.confnet.bridge.bridge-nf-call-iptables=1
net.bridge.bridge-nf-call-ip6tables=1
net.ipv4.ip_forward=1
vm.swappiness=0
vm.overcommit_memory=1
vm.panic_on_oom=0
fs.inotify.max_user_watches=89100

repo文件kubernetes.repo

# vim roles/docker_install/templates/kubernetes.repo[kubernetes]name=Kubernetes
baseurl=http://mirrors.aliyun.com/kubernetes/yum/repos/kubernetes-el7-x86_64
enabled=1
gpgcheck=0
repo_gpgcheck=0
gpgkey=http://mirrors.aliyun.com/kubernetes/yum/doc/yum-key.gpg
       http://mirrors.aliyun.com/kubernetes/yum/doc/rpm-package-key.gpg

  • 环境准备prepare.yml:
# vim roles/docker_install/tasks/prepare.yml

- name: 关闭firewalld 
  service: name=firewalld state=stopped enabled=no  
- name: 临时关闭 selinux  shell: "setenforce 0"
  failed_when: false- name: 永久关闭 selinux  lineinfile:
    dest: /etc/selinux/config    regexp: "^SELINUX="
    line: "SELINUX=disabled"- name: 添加EPEL仓库  yum: name=epel-release state=latest- name: 安装常用软件包  yum:
    name:
      - vim      - lrzsz      - net-tools      - wget      - curl      - bash-completion      - rsync      - gcc      - unzip      - git      - iptables      - conntrack      - ipvsadm      - ipset      - jq      - sysstat      - libseccomp    state: latest- name: 更新系统  shell: "yum update -y --exclude kubeadm,kubelet,kubectl"
  ignore_errors: yes  args:
    warn: False
    - name: 配置iptables  shell: "iptables -F && iptables -X && iptables -F -t nat && iptables -X -t nat && iptables -P FORWARD ACCEPT"- name: 关闭swap  shell: "swapoff -a && sed -i '/swap/s/^\\(.*\\)$/#\\1/g' /etc/fstab"
  - name: 系统配置  template: src=kubernetes.conf dest=/etc/sysctl.d/kubernetes.conf- name: 加载br_netfilter  shell: "modprobe br_netfilter"- name: 生效配置  shell: "sysctl -p /etc/sysctl.d/kubernetes.conf"

  • docker安装install.yml:
# vim roles/docker_install/tasks/install.yml

- name: 创建software目录  file: name={{ SOURCE_DIR }} state=directory- name: 更改hostname  raw: "echo {{ hostname }} > /etc/hostname"- name: 更改生效  shell: "hostname {{ hostname }}"- name: 设置本地dns  shell: "if [ `grep '{{ ansible_ssh_host }} {{ hostname }}' /etc/hosts |wc -l` -eq 0 ]; then echo {{ ansible_ssh_host }} {{ hostname }} >> /etc/hosts; fi"- name: 下载repo文件  shell: "if [ ! -f /etc/yum.repos.d/docker.repo ]; then curl http://mirrors.aliyun.com/docker-ce/linux/centos/docker-ce.repo -o /etc/yum.repos.d/docker.repo; fi"- name: 生成缓存  shell: "yum makecache fast"
  args:
    warn: False- name: 安装docker-ce  yum: 
    name: docker-ce    state: latest- name: 启动docker并开机启动  service:
    name: docker    state: started    enabled: yes    
- name: 配置docker  template: src=daemon.json dest=/etc/docker/daemon.json- name: 重启docker  service:
    name: docker    state: restarted    
- name: 配置kubernetes源  template: src=kubernetes.repo dest=/etc/yum.repos.d/kubernetes.repo- name: 安装kubernetes-cni  yum: 
    name: kubernetes-cni    state: latest  

- name: 安装kubeadm、kubelet、kubectl  shell: "yum install -y kubeadm-{{ VERSION }} kubelet-{{ VERSION }} kubectl-{{ VERSION }} --disableexcludes=kubernetes"
  args:
    warn: False- name: 启动kubelet并开机启动  service:
    name: kubelet    state: started    enabled: yes

  • 引用文件main.yml:
# vim roles/docker_install/tasks/main.yml#引用prepare、install模块- include: prepare.yml
- include: install.yml


init部分

  • 创建init入口文件,用来调用init_install:
# vim init.yml#用于初始化k8s集群- hosts: master
  remote_user: root
  gather_facts: True

  roles:
    - init_install

  • 创建变量:
# vim roles/init_install/vars/main.yml#定义初始化k8s集群中的变量#kubernetes版本VERSION: v1.14.0#Pod网段POD_CIDR: 172.10.0.0/16#master虚拟ip(建议为同网段地址)MASTER_VIP: 192.168.30.188#keepalived用到的网卡接口名VIP_IF: ens33

SOURCE_DIR: /software
Cluster_Num: "{{ groups['all'] | length }}"Virtual_Router_ID: 68

  • 创建模板文件:

keepalived master配置文件 keepalived-master.conf

# vim roles/init_install/templates/keepalived-master.conf! Configuration File for keepalived
global_defs {
    router_id keepalive-master}vrrp_script check_apiserver {
    script "/etc/keepalived/check-apiserver.sh"
    interval 3
    weight -{{ Cluster_Num }}}vrrp_instance VI-kube-master {
    state MASTER
    interface {{ VIP_IF }}
    virtual_router_id {{ Virtual_Router_ID }}
    priority 100
    dont_track_primary
    advert_int 3
    virtual_ipaddress {
        {{ MASTER_VIP }}
    }
    track_script {
        check_apiserver    }}

keepalived检查脚本 check-apiserver.sh

# vim roles/init_install/templates/check-apiserver.sh#!/bin/shnetstat -lntp |grep 6443 || exit 1

kubeadm配置文件 kubeadm-config.yaml

# vim roles/init_install/templates/kubeadm-config.yamlapiVersion: kubeadm.k8s.io/v1beta1
kind: ClusterConfiguration
kubernetesVersion: {{ VERSION }}controlPlaneEndpoint: "{{ MASTER_VIP }}:6443"networking:    # This CIDR is a Calico default. Substitute or remove for your CNI provider.
    podSubnet: "{{ POD_CIDR }}"imageRepository: registry.cn-hangzhou.aliyuncs.com/imooc

  • 集群初始化install.yml:
# vim roles/init_install/tasks/install.yml

- name: 安装keepalived  yum: name=keepalived state=present- name: 拷贝keepalived配置文件  template: src=keepalived-master.conf dest=/etc/keepalived/keepalived.conf  
- name: 拷贝keepalived检查脚本  template: src=check-apiserver.sh dest=/etc/keepalived/check-apiserver.sh mode=0755- name: 启动keepalived并开机启动  service:
    name: keepalived    state: started    enabled: yes- name: 拷贝kubeadm配置文件  template: src=kubeadm-config.yaml dest={{ SOURCE_DIR }}- name: 集群初始化准备1  shell: "swapoff -a && kubeadm reset -f"- name: 集群初始化准备2  shell: "systemctl daemon-reload && systemctl restart kubelet"
  - name: 集群初始化准备3  shell: "iptables -F && iptables -t nat -F && iptables -t mangle -F && iptables -X"- name: 拉取kube-scheduler镜像  shell: "docker pull registry.cn-hangzhou.aliyuncs.com/imooc/kube-scheduler:{{ VERSION }}"- name: 集群初始化  shell: "kubeadm init --config={{ SOURCE_DIR }}/kubeadm-config.yaml --experimental-upload-certs &>{{ SOURCE_DIR }}/token.txt"- name: 获取master的token  shell: "grep -B2 'experimental-control-plane' {{ SOURCE_DIR }}/token.txt > {{ SOURCE_DIR }}/master.sh"- name: 获取node的token  shell: "grep -A1 'kubeadm join' {{ SOURCE_DIR }}/token.txt |tail -2 > {{ SOURCE_DIR }}/node.sh"- name: 分发master.sh  shell: "ansible all -m copy -a 'src={{ SOURCE_DIR }}/master.sh dest={{ SOURCE_DIR }} mode=0755'"
  args:
    warn: False
    - name: 分发node.sh  shell: "ansible all -m copy -a 'src={{ SOURCE_DIR }}/node.sh dest={{ SOURCE_DIR }} mode=0755'"
  args:
    warn: False- name: 创建 $HOME/.kube 目录  file: name=$HOME/.kube state=directory  
- name: 拷贝KubeConfig  copy: src=/etc/kubernetes/admin.conf dest=$HOME/.kube/config owner=root group=root- name: kubectl命令补全1  shell: "kubectl completion bash > $HOME/.kube/completion.bash.inc"
 - name: kubectl命令补全2  shell: "if [ `grep 'source $HOME/.kube/completion.bash.inc' $HOME/.bash_profile |wc -l` -eq 0 ]; then echo 'source $HOME/.kube/completion.bash.inc' >> $HOME/.bash_profile; fi"
  - name: 生效配置  shell: "source $HOME/.bash_profile"
  ignore_errors: yes

  • 引用文件main.yml:
# vim roles/init_install/tasks/main.yml#引用install模块- include: install.yml


master部分

  • 创建master入口文件,用来调用master_install:
# vim master.yml#用于集群新增master- hosts: add_master
  remote_user: root
  gather_facts: True

  roles:
    - master_install

  • 创建变量:
# vim roles/master_install/vars/main.yml#定义新增master到集群中的变量#注意与keepalived master一致#master虚拟ip(建议为同网段地址)MASTER_VIP: 192.168.30.188#keepalived用到的网卡接口名VIP_IF: ens33

SOURCE_DIR: /software
Cluster_Num: "{{ groups['all'] | length }}"Virtual_Router_ID: 68

  • 创建模板文件:

keepalived backup配置文件 keepalived-backup.conf

# vim roles/master_install/templates/keepalived-backup.conf! Configuration File for keepalived
global_defs {
    router_id keepalive-backup}vrrp_script check_apiserver {
    script "/etc/keepalived/check-apiserver.sh"
    interval 3
    weight -{{ Cluster_Num }}}vrrp_instance VI-kube-master {
    state BACKUP
    interface {{ VIP_IF }}
    virtual_router_id {{ Virtual_Router_ID }}
    priority 99
    dont_track_primary
    advert_int 3
    virtual_ipaddress {
        {{ MASTER_VIP }}
    }
    track_script {
        check_apiserver    }}

keepalived检查脚本 check-apiserver.sh

# vim roles/master_install/templates/check-apiserver.sh#!/bin/shnetstat -lntp |grep 6443 || exit 1

  • 添加master到集群install.yml:
# vim roles/master_install/tasks/install.yml

- name: 安装keepalived  yum: name=keepalived state=present- name: 拷贝keepalived配置文件  template: src=keepalived-backup.conf dest=/etc/keepalived/keepalived.conf  
- name: 拷贝keepalived检查脚本  template: src=check-apiserver.sh dest=/etc/keepalived/check-apiserver.sh mode=0755- name: 启动keepalived并开机启动  service:
    name: keepalived    state: started    enabled: yes- name: 集群初始化准备1  shell: "swapoff -a && kubeadm reset -f"- name: 集群初始化准备2  shell: "systemctl daemon-reload && systemctl restart kubelet"
  - name: 集群初始化准备3  shell: "iptables -F && iptables -t nat -F && iptables -t mangle -F && iptables -X"
  - name: 集群增加master  script: "{{ SOURCE_DIR }}/master.sh"
  - name: 创建 $HOME/.kube 目录  file: name=$HOME/.kube state=directory  
- name: 拷贝KubeConfig  copy: src=/etc/kubernetes/admin.conf dest=$HOME/.kube/config owner=root group=root- name: kubectl命令补全1  shell: "kubectl completion bash > $HOME/.kube/completion.bash.inc"
 - name: kubectl命令补全2  shell: "if [ `grep 'source $HOME/.kube/completion.bash.inc' $HOME/.bash_profile |wc -l` -eq 0 ]; then echo 'source $HOME/.kube/completion.bash.inc' >> $HOME/.bash_profile; fi"
  - name: 生效配置  shell: "source $HOME/.bash_profile"
  ignore_errors: yes  
- name: 删除master的token  file: name={{ SOURCE_DIR }}/master.sh state=absent- name: 删除node的token  file: name={{ SOURCE_DIR }}/node.sh state=absent

  • 引用文件main.yml:
# vim roles/master_install/tasks/main.yml#引用install模块- include: install.yml


node部分

  • 创建node入口文件,用来调用node_install:
# vim node.yml#用于集群增加node- hosts: add_node
  remote_user: root
  gather_facts: True

  roles:
    - node_install

  • 创建变量:
# vim roles/node_install/vars/main.yml#定义新增node到集群中的变量SOURCE_DIR: /software

  • 添加node到集群install.yml:
# vim roles/node_install/tasks/install.yml

- name: 集群初始化准备1  shell: "swapoff -a && kubeadm reset -f"- name: 集群初始化准备2  shell: "systemctl daemon-reload && systemctl restart kubelet"
  - name: 集群初始化准备3  shell: "iptables -F && iptables -t nat -F && iptables -t mangle -F && iptables -X"
  - name: 集群增加node  script: "{{ SOURCE_DIR }}/node.sh"- name: 删除master的token  file: name={{ SOURCE_DIR }}/master.sh state=absent- name: 删除node的token  file: name={{ SOURCE_DIR }}/node.sh state=absent

  • 引用文件main.yml:
# vim roles/node_install/tasks/main.yml#引用install模块- include: install.yml


addons部分

  • 创建addons入口文件,用来调用addons_install:
# vim addons.yml#用于安装k8s集群插件- hosts: master
  remote_user: root
  gather_facts: True

  roles:
    - addons_install

  • 创建变量:
# vim roles/addons_install/vars/main.yml#定义k8s集群插件安装中的变量#Pod网段POD_CIDR: 172.10.0.0/16

SOURCE_DIR: /software

  • 创建模板文件:

calico rbac配置文件 calico-rbac-kdd.yaml

# vim roles/addons_install/templates/calico-rbac-kdd.yaml

# Calico Version v3.1.3# https://docs.projectcalico.org/v3.1/releases#v3.1.3kind: ClusterRoleapiVersion: rbac.authorization.k8s.io/v1beta1metadata:
  name: calico-noderules:
  - apiGroups: [""]
    resources:
      - namespaces    verbs:
      - get      - list      - watch  - apiGroups: [""]
    resources:
      - pods/status    verbs:
      - update  - apiGroups: [""]
    resources:
      - pods    verbs:
      - get      - list      - watch      - patch  - apiGroups: [""]
    resources:
      - services    verbs:
      - get  - apiGroups: [""]
    resources:
      - endpoints    verbs:
      - get  - apiGroups: [""]
    resources:
      - nodes    verbs:
      - get      - list      - update      - watch  - apiGroups: ["extensions"]
    resources:
      - networkpolicies    verbs:
      - get      - list      - watch  - apiGroups: ["networking.k8s.io"]
    resources:
      - networkpolicies    verbs:
      - watch      - list  - apiGroups: ["crd.projectcalico.org"]
    resources:
      - globalfelixconfigs      - felixconfigurations      - bgppeers      - globalbgpconfigs      - bgpconfigurations      - ippools      - globalnetworkpolicies      - globalnetworksets      - networkpolicies      - clusterinformations      - hostendpoints    verbs:
      - create      - get      - list      - update      - watch---apiVersion: rbac.authorization.k8s.io/v1beta1kind: ClusterRoleBindingmetadata:
  name: calico-noderoleRef:
  apiGroup: rbac.authorization.k8s.io  kind: ClusterRole  name: calico-nodesubjects:- kind: ServiceAccount  name: calico-node  namespace: kube-system

calico配置文件 calico.yaml

# vim roles/addons_install/templates/calico.yaml

# Calico Version v3.1.3# https://docs.projectcalico.org/v3.1/releases#v3.1.3# This manifest includes the following component versions:#   calico/node:v3.1.3#   calico/cni:v3.1.3# This ConfigMap is used to configure a self-hosted Calico installation.kind: ConfigMapapiVersion: v1metadata:
  name: calico-config  namespace: kube-systemdata:
  # To enable Typha, set this to "calico-typha" *and* set a non-zero value for Typha replicas
  # below.  We recommend using Typha if you have more than 50 nodes. Above 100 nodes it is
  # essential.
  typha_service_name: "calico-typha"

  # The CNI network configuration to install on each node.
  cni_network_config: |-
    {
      "name": "k8s-pod-network",
      "cniVersion": "0.3.0",
      "plugins": [
        {
          "type": "calico",
          "log_level": "info",
          "datastore_type": "kubernetes",
          "nodename": "__KUBERNETES_NODE_NAME__",
          "mtu": 1500,
          "ipam": {
            "type": "host-local",
            "subnet": "usePodCidr"
          },
          "policy": {
            "type": "k8s"
          },
          "kubernetes": {
            "kubeconfig": "__KUBECONFIG_FILEPATH__"
          }
        },
        {
          "type": "portmap",
          "snat": true,
          "capabilities": {"portMappings": true}
        }
      ]
    }---# This manifest creates a Service, which will be backed by Calico's Typha daemon.# Typha sits in between Felix and the API server, reducing Calico's load on the API server.apiVersion: v1kind: Servicemetadata:
  name: calico-typha  namespace: kube-system  labels:
    k8s-app: calico-typhaspec:
  ports:
    - port: 5473
      protocol: TCP      targetPort: calico-typha      name: calico-typha  selector:
    k8s-app: calico-typha---# This manifest creates a Deployment of Typha to back the above service.apiVersion: apps/v1beta1kind: Deploymentmetadata:
  name: calico-typha  namespace: kube-system  labels:
    k8s-app: calico-typhaspec:
  # Number of Typha replicas.  To enable Typha, set this to a non-zero value *and* set the
  # typha_service_name variable in the calico-config ConfigMap above.
  #
  # We recommend using Typha if you have more than 50 nodes.  Above 100 nodes it is essential
  # (when using the Kubernetes datastore).  Use one replica for every 100-200 nodes.  In
  # production, we recommend running at least 3 replicas to reduce the impact of rolling upgrade.
  replicas: 1
  revisionHistoryLimit: 2
  template:
    metadata:
      labels:
        k8s-app: calico-typha      annotations:
        # This, along with the CriticalAddonsOnly toleration below, marks the pod as a critical
        # add-on, ensuring it gets priority scheduling and that its resources are reserved
        # if it ever gets evicted.
        scheduler.alpha.kubernetes.io/critical-pod: ''
    spec:
      hostNetwork: true
      tolerations:
        # Mark the pod as a critical add-on for rescheduling.
        - key: CriticalAddonsOnly          operator: Exists      # Since Calico can't network a pod until Typha is up, we need to run Typha itself
      # as a host-networked pod.
      serviceAccountName: calico-node      containers:
      - image: quay.io/calico/typha:v0.7.4        name: calico-typha        ports:
        - containerPort: 5473
          name: calico-typha          protocol: TCP        env:
          # Enable "info" logging by default.  Can be set to "debug" to increase verbosity.
          - name: TYPHA_LOGSEVERITYSCREEN            value: "info"
          # Disable logging to file and syslog since those don't make sense in Kubernetes.
          - name: TYPHA_LOGFILEPATH            value: "none"
          - name: TYPHA_LOGSEVERITYSYS            value: "none"
          # Monitor the Kubernetes API to find the number of running instances and rebalance
          # connections.
          - name: TYPHA_CONNECTIONREBALANCINGMODE            value: "kubernetes"
          - name: TYPHA_DATASTORETYPE            value: "kubernetes"
          - name: TYPHA_HEALTHENABLED            value: "true"
          # Uncomment these lines to enable prometheus metrics.  Since Typha is host-networked,
          # this opens a port on the host, which may need to be secured.
          #- name: TYPHA_PROMETHEUSMETRICSENABLED
          #  value: "true"
          #- name: TYPHA_PROMETHEUSMETRICSPORT
          #  value: "9093"
        livenessProbe:
          httpGet:
            path: /liveness            port: 9098
          periodSeconds: 30
          initialDelaySeconds: 30
        readinessProbe:
          httpGet:
            path: /readiness            port: 9098
          periodSeconds: 10---# This manifest installs the calico/node container, as well# as the Calico CNI plugins and network config on# each master and worker node in a Kubernetes cluster.kind: DaemonSetapiVersion: extensions/v1beta1metadata:
  name: calico-node  namespace: kube-system  labels:
    k8s-app: calico-nodespec:
  selector:
    matchLabels:
      k8s-app: calico-node  updateStrategy:
    type: RollingUpdate    rollingUpdate:
      maxUnavailable: 1
  template:
    metadata:
      labels:
        k8s-app: calico-node      annotations:
        # This, along with the CriticalAddonsOnly toleration below,
        # marks the pod as a critical add-on, ensuring it gets
        # priority scheduling and that its resources are reserved
        # if it ever gets evicted.
        scheduler.alpha.kubernetes.io/critical-pod: ''
    spec:
      hostNetwork: true
      tolerations:
        # Make sure calico/node gets scheduled on all nodes.
        - effect: NoSchedule          operator: Exists        # Mark the pod as a critical add-on for rescheduling.
        - key: CriticalAddonsOnly          operator: Exists        - effect: NoExecute          operator: Exists      serviceAccountName: calico-node      # Minimize downtime during a rolling upgrade or deletion; tell Kubernetes to do a "force
      # deletion": https://kubernetes.io/docs/concepts/workloads/pods/pod/#termination-of-pods.
      terminationGracePeriodSeconds: 0
      containers:
        # Runs calico/node container on each Kubernetes node.  This
        # container programs network policy and routes on each
        # host.
        - name: calico-node          image: quay.io/calico/node:v3.1.3          env:
            # Use Kubernetes API as the backing datastore.
            - name: DATASTORE_TYPE              value: "kubernetes"
            # Enable felix info logging.
            - name: FELIX_LOGSEVERITYSCREEN              value: "info"
            # Cluster type to identify the deployment type
            - name: CLUSTER_TYPE              value: "k8s,bgp"
            # Disable file logging so `kubectl logs` works.
            - name: CALICO_DISABLE_FILE_LOGGING              value: "true"
            # Set Felix endpoint to host default action to ACCEPT.
            - name: FELIX_DEFAULTENDPOINTTOHOSTACTION              value: "ACCEPT"
            # Disable IPV6 on Kubernetes.
            - name: FELIX_IPV6SUPPORT              value: "false"
            # Set MTU for tunnel device used if ipip is enabled
            - name: FELIX_IPINIPMTU              value: "1440"
            # Wait for the datastore.
            - name: WAIT_FOR_DATASTORE              value: "true"
            # The default IPv4 pool to create on startup if none exists. Pod IPs will be
            # chosen from this range. Changing this value after installation will have
            # no effect. This should fall within `--cluster-cidr`.
            - name: CALICO_IPV4POOL_CIDR              value: "{{ POD_CIDR }}"
            # Enable IPIP
            - name: CALICO_IPV4POOL_IPIP              value: "Always"
            # Enable IP-in-IP within Felix.
            - name: FELIX_IPINIPENABLED              value: "true"
            # Typha support: controlled by the ConfigMap.
            - name: FELIX_TYPHAK8SSERVICENAME              valueFrom:
                configMapKeyRef:
                  name: calico-config                  key: typha_service_name            # Set based on the k8s node name.
            - name: NODENAME              valueFrom:
                fieldRef:
                  fieldPath: spec.nodeName            # Auto-detect the BGP IP address.
            - name: IP              value: "autodetect"
            - name: FELIX_HEALTHENABLED              value: "true"
          securityContext:
            privileged: true
          resources:
            requests:
              cpu: 250m          livenessProbe:
            httpGet:
              path: /liveness              port: 9099
            periodSeconds: 10
            initialDelaySeconds: 10
            failureThreshold: 6
          readinessProbe:
            httpGet:
              path: /readiness              port: 9099
            periodSeconds: 10
          volumeMounts:
            - mountPath: /lib/modules              name: lib-modules              readOnly: true
            - mountPath: /var/run/calico              name: var-run-calico              readOnly: false
            - mountPath: /var/lib/calico              name: var-lib-calico              readOnly: false
        # This container installs the Calico CNI binaries
        # and CNI network config file on each node.
        - name: install-cni          image: quay.io/calico/cni:v3.1.3          command: ["/install-cni.sh"]
          env:
            # Name of the CNI config file to create.
            - name: CNI_CONF_NAME              value: "10-calico.conflist"
            # The CNI network config to install on each node.
            - name: CNI_NETWORK_CONFIG              valueFrom:
                configMapKeyRef:
                  name: calico-config                  key: cni_network_config            # Set the hostname based on the k8s node name.
            - name: KUBERNETES_NODE_NAME              valueFrom:
                fieldRef:
                  fieldPath: spec.nodeName          volumeMounts:
            - mountPath: /host/opt/cni/bin              name: cni-bin-dir            - mountPath: /host/etc/cni/net.d              name: cni-net-dir      volumes:
        # Used by calico/node.
        - name: lib-modules          hostPath:
            path: /lib/modules        - name: var-run-calico          hostPath:
            path: /var/run/calico        - name: var-lib-calico          hostPath:
            path: /var/lib/calico        # Used to install CNI.
        - name: cni-bin-dir          hostPath:
            path: /opt/cni/bin        - name: cni-net-dir          hostPath:
            path: /etc/cni/net.d# Create all the CustomResourceDefinitions needed for# Calico policy and networking mode.---apiVersion: apiextensions.k8s.io/v1beta1kind: CustomResourceDefinitionmetadata:
   name: felixconfigurations.crd.projectcalico.orgspec:
  scope: Cluster  group: crd.projectcalico.org  version: v1  names:
    kind: FelixConfiguration    plural: felixconfigurations    singular: felixconfiguration---apiVersion: apiextensions.k8s.io/v1beta1kind: CustomResourceDefinitionmetadata:
  name: bgppeers.crd.projectcalico.orgspec:
  scope: Cluster  group: crd.projectcalico.org  version: v1  names:
    kind: BGPPeer    plural: bgppeers    singular: bgppeer---apiVersion: apiextensions.k8s.io/v1beta1kind: CustomResourceDefinitionmetadata:
  name: bgpconfigurations.crd.projectcalico.orgspec:
  scope: Cluster  group: crd.projectcalico.org  version: v1  names:
    kind: BGPConfiguration    plural: bgpconfigurations    singular: bgpconfiguration---apiVersion: apiextensions.k8s.io/v1beta1kind: CustomResourceDefinitionmetadata:
  name: ippools.crd.projectcalico.orgspec:
  scope: Cluster  group: crd.projectcalico.org  version: v1  names:
    kind: IPPool    plural: ippools    singular: ippool---apiVersion: apiextensions.k8s.io/v1beta1kind: CustomResourceDefinitionmetadata:
  name: hostendpoints.crd.projectcalico.orgspec:
  scope: Cluster  group: crd.projectcalico.org  version: v1  names:
    kind: HostEndpoint    plural: hostendpoints    singular: hostendpoint---apiVersion: apiextensions.k8s.io/v1beta1kind: CustomResourceDefinitionmetadata:
  name: clusterinformations.crd.projectcalico.orgspec:
  scope: Cluster  group: crd.projectcalico.org  version: v1  names:
    kind: ClusterInformation    plural: clusterinformations    singular: clusterinformation---apiVersion: apiextensions.k8s.io/v1beta1kind: CustomResourceDefinitionmetadata:
  name: globalnetworkpolicies.crd.projectcalico.orgspec:
  scope: Cluster  group: crd.projectcalico.org  version: v1  names:
    kind: GlobalNetworkPolicy    plural: globalnetworkpolicies    singular: globalnetworkpolicy---apiVersion: apiextensions.k8s.io/v1beta1kind: CustomResourceDefinitionmetadata:
  name: globalnetworksets.crd.projectcalico.orgspec:
  scope: Cluster  group: crd.projectcalico.org  version: v1  names:
    kind: GlobalNetworkSet    plural: globalnetworksets    singular: globalnetworkset---apiVersion: apiextensions.k8s.io/v1beta1kind: CustomResourceDefinitionmetadata:
  name: networkpolicies.crd.projectcalico.orgspec:
  scope: Namespaced  group: crd.projectcalico.org  version: v1  names:
    kind: NetworkPolicy    plural: networkpolicies    singular: networkpolicy---apiVersion: v1kind: ServiceAccountmetadata:
  name: calico-node  namespace: kube-system

dashboard配置文件 dashboard-all.yaml

# vim roles/addons_install/templates/dashboard-all.yaml

apiVersion: v1kind: ConfigMapmetadata:
  labels:
    k8s-app: kubernetes-dashboard    # Allows editing resource and makes sure it is created first.
    addonmanager.kubernetes.io/mode: EnsureExists  name: kubernetes-dashboard-settings  namespace: kube-system---apiVersion: v1kind: ServiceAccountmetadata:
  labels:
    k8s-app: kubernetes-dashboard    addonmanager.kubernetes.io/mode: Reconcile  name: kubernetes-dashboard  namespace: kube-system---apiVersion: apps/v1kind: Deploymentmetadata:
  name: kubernetes-dashboard  namespace: kube-system  labels:
    k8s-app: kubernetes-dashboard    kubernetes.io/cluster-service: "true"
    addonmanager.kubernetes.io/mode: Reconcilespec:
  selector:
    matchLabels:
      k8s-app: kubernetes-dashboard  template:
    metadata:
      labels:
        k8s-app: kubernetes-dashboard      annotations:
        scheduler.alpha.kubernetes.io/critical-pod: ''
        seccomp.security.alpha.kubernetes.io/pod: 'docker/default'
    spec:
      priorityClassName: system-cluster-critical      containers:
      - name: kubernetes-dashboard        image: registry.cn-hangzhou.aliyuncs.com/imooc/kubernetes-dashboard-amd64:v1.8.3        resources:
          limits:
            cpu: 100m            memory: 300Mi          requests:
            cpu: 50m            memory: 100Mi        ports:
        - containerPort: 8443
          protocol: TCP        args:
          # PLATFORM-SPECIFIC ARGS HERE
          - --auto-generate-certificates        volumeMounts:
        - name: kubernetes-dashboard-certs          mountPath: /certs        - name: tmp-volume          mountPath: /tmp        livenessProbe:
          httpGet:
            scheme: HTTPS            path: /            port: 8443
          initialDelaySeconds: 30
          timeoutSeconds: 30
      volumes:
      - name: kubernetes-dashboard-certs        secret:
          secretName: kubernetes-dashboard-certs      - name: tmp-volume        emptyDir: {}
      serviceAccountName: kubernetes-dashboard      tolerations:
      - key: "CriticalAddonsOnly"
        operator: "Exists"---apiVersion: rbac.authorization.k8s.io/v1kind: Rolemetadata:
  labels:
    k8s-app: kubernetes-dashboard    addonmanager.kubernetes.io/mode: Reconcile  name: kubernetes-dashboard-minimal  namespace: kube-systemrules:
  # Allow Dashboard to get, update and delete Dashboard exclusive secrets.- apiGroups: [""]
  resources: ["secrets"]
  resourceNames: ["kubernetes-dashboard-key-holder", "kubernetes-dashboard-certs"]
  verbs: ["get", "update", "delete"]
  # Allow Dashboard to get and update 'kubernetes-dashboard-settings' config map.- apiGroups: [""]
  resources: ["configmaps"]
  resourceNames: ["kubernetes-dashboard-settings"]
  verbs: ["get", "update"]
  # Allow Dashboard to get metrics from heapster.- apiGroups: [""]
  resources: ["services"]
  resourceNames: ["heapster"]
  verbs: ["proxy"]- apiGroups: [""]
  resources: ["services/proxy"]
  resourceNames: ["heapster", "http:heapster:", "https:heapster:"]
  verbs: ["get"]---apiVersion: rbac.authorization.k8s.io/v1kind: RoleBindingmetadata:
  name: kubernetes-dashboard-minimal  namespace: kube-system  labels:
    k8s-app: kubernetes-dashboard    addonmanager.kubernetes.io/mode: ReconcileroleRef:
  apiGroup: rbac.authorization.k8s.io  kind: Role  name: kubernetes-dashboard-minimalsubjects:- kind: ServiceAccount  name: kubernetes-dashboard  namespace: kube-system---apiVersion: v1kind: Secretmetadata:
  labels:
    k8s-app: kubernetes-dashboard    # Allows editing resource and makes sure it is created first.
    addonmanager.kubernetes.io/mode: EnsureExists  name: kubernetes-dashboard-certs  namespace: kube-systemtype: Opaque---apiVersion: v1kind: Secretmetadata:
  labels:
    k8s-app: kubernetes-dashboard    # Allows editing resource and makes sure it is created first.
    addonmanager.kubernetes.io/mode: EnsureExists  name: kubernetes-dashboard-key-holder  namespace: kube-systemtype: Opaque---apiVersion: v1kind: Servicemetadata:
  name: kubernetes-dashboard  namespace: kube-system  labels:
    k8s-app: kubernetes-dashboard    kubernetes.io/cluster-service: "true"
    addonmanager.kubernetes.io/mode: Reconcilespec:
  selector:
    k8s-app: kubernetes-dashboard  ports:
  - port: 443
    targetPort: 8443
    nodePort: 30005
  type: NodePort

  • coredns安装calico.yml:
# vim roles/addons_install/tasks/calico.yml

- name: 创建addons目录  file: name=/etc/kubernetes/addons state=directory- name: 拷贝calico-rbac-kdd.yaml  template: src=calico-rbac-kdd.yaml dest=/etc/kubernetes/addons- name: 拷贝calico.yaml  template: src=calico.yaml dest=/etc/kubernetes/addons- name: 拉取calico typha镜像  shell: "ansible all -m shell -a 'docker pull registry.cn-hangzhou.aliyuncs.com/liuyi01/calico-typha:v0.7.4'"- name: tag calico typha镜像  shell: "ansible all -m shell -a 'docker tag registry.cn-hangzhou.aliyuncs.com/liuyi01/calico-typha:v0.7.4 quay.io/calico/typha:v0.7.4'"
  - name: 拉取calico node镜像  shell: "ansible all -m shell -a 'docker pull registry.cn-hangzhou.aliyuncs.com/liuyi01/calico-node:v3.1.3'"- name: tag calico node镜像  shell: "ansible all -m shell -a 'docker tag registry.cn-hangzhou.aliyuncs.com/liuyi01/calico-node:v3.1.3 quay.io/calico/node:v3.1.3'"- name: 拉取calico cni镜像  shell: "ansible all -m shell -a 'docker pull registry.cn-hangzhou.aliyuncs.com/liuyi01/calico-cni:v3.1.3'"
  - name: tag calico cni镜像  shell: "ansible all -m shell -a 'docker tag registry.cn-hangzhou.aliyuncs.com/liuyi01/calico-cni:v3.1.3 quay.io/calico/cni:v3.1.3'"
  - name: 创建calico-rbac  shell: "kubectl apply -f /etc/kubernetes/addons/calico-rbac-kdd.yaml"- name: 部署calico  shell: "kubectl apply -f /etc/kubernetes/addons/calico.yaml"

  • dashboard安装dashboard.yml:
# vim roles/addons_install/tasks/dashboard.yml

- name: 拷贝dashboard-all.yaml  template: src=dashboard-all.yaml dest=/etc/kubernetes/addons- name: 拉取dashboard镜像  shell: "ansible all -m shell -a 'docker pull registry.cn-hangzhou.aliyuncs.com/imooc/kubernetes-dashboard-amd64:v1.8.3'"- name: 部署dashboard  shell: "kubectl apply -f /etc/kubernetes/addons/dashboard-all.yaml"
  - name: 创建ServiceaAccount  shell: "kubectl create sa dashboard-admin -n kube-system"- name: 权限绑定  shell: "kubectl create clusterrolebinding dashboard-admin --clusterrole=cluster-admin --serviceaccount=kube-system:dashboard-admin"- name: 获取登录token  shell: "ADMIN_SECRET=$(kubectl get secrets -n kube-system | grep dashboard-admin | awk '{print $1}'); kubectl describe secret -n kube-system ${ADMIN_SECRET} | grep -E '^token' | awk '{print $2}' > {{ SOURCE_DIR }}/token.txt"
  register: token- name: 显示token位置  debug: var=token.cmd verbosity=0

  • 引用文件main.yml:
# vim roles/addons_install/tasks/main.yml

#引用calico、dashboard模块- include: calico.yml- include: dashboard.yml


安装测试

  • 执行安装:
# ansible-playbook k8s.yml

# kubectl get nodesNAME      STATUS   ROLES    AGE     VERSION
master1   Ready    master   8m30s   v1.14.0
master2   Ready    master   6m38s   v1.14.0
node1     Ready    <none>   5m50s   v1.14.0
node2     Ready    <none>   5m49s   v1.14.0
node3     Ready    <none>   5m49s   v1.14.0# kubectl get pods -n kube-systemNAME                                   READY   STATUS    RESTARTS   AGE
calico-node-9jqrh                      2/2     Running   0          19h
calico-node-brfb8                      2/2     Running   0          19h
calico-node-ncjxp                      2/2     Running   0          19h
calico-node-qsffm                      2/2     Running   0          19h
calico-node-smwq5                      2/2     Running   0          19h
calico-typha-666749994b-h99rr          1/1     Running   0          19h
coredns-8567978547-6fvmp               1/1     Running   1          19h
coredns-8567978547-q2rz5               1/1     Running   1          19h
etcd-master1                           1/1     Running   0          19h
etcd-master2                           1/1     Running   0          19h
kube-apiserver-master1                 1/1     Running   0          19h
kube-apiserver-master2                 1/1     Running   0          19h
kube-controller-manager-master1        1/1     Running   0          19h
kube-controller-manager-master2        1/1     Running   0          19h
kube-proxy-2wfnz                       1/1     Running   0          19h
kube-proxy-46gtf                       1/1     Running   0          19h
kube-proxy-mg87t                       1/1     Running   0          19h
kube-proxy-s5f5j                       1/1     Running   0          19h
kube-proxy-vqdhx                       1/1     Running   0          19h
kube-scheduler-master1                 1/1     Running   0          19h
kube-scheduler-master2                 1/1     Running   0          19h
kubernetes-dashboard-5bd4bfc87-rkvfh   1/1     Running   0          19h

使用火狐浏览器访问https://192.168.30.128:30005

在这里插入图片描述

# cat /software/token.txteyJhbGciOiJSUzI1NiIsImtpZCI6IiJ9.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.xOSS0owznhpdN0DmALG04ehP7Xw-qYEjX8YQAHy8vA017UFvRxd8oH4EsfkGEkzLWC3kwWUWCxRs7yEVUgGzx7Rp2lJ5xceyga448Kffj9GOQapAF8J_2SP9j2tCll465GSKtjODJgpwnknJ8YqMt6mn-jH9Cn49ljf6rSHuyk2_2qf_PX2ioIKHTKyFLMD-6ci-tUpWHOEQdKlxi3K7LJek4qlJ04hKckcphHZf35YEnBDcB2uvWsW7P9k_2GIwXwjoPLAlidmk56WGtrCy9erZFg1W1aeVh7z9aqECKSs0Jnt-bafPaBwbqpjbkjhGKC8vznBslxO2STaiR0-j1A

在这里插入图片描述

测试安装没有问题,注意kubernetes组件版本尽量一致。已存放至个人gitgub:ansible-playbook


这篇关于ansible-playbook部署K8S高可用集群的文章就介绍到这儿,希望我们推荐的文章对大家有所帮助,也希望大家多多支持为之网!