一、利用ansible部署kubernetes集群环境准备
基于二进制方式部署和利用ansible-playbook实现自动化：既提供一键安装脚本，也可以分步执行安装各个组件，同时讲解每一步主要参数配置和注意事项；二进制方式部署有助于理解系统各组件的交互原理和熟悉组件启动参数，有助于快速排查解决实际问题。
版本组件：
kubernetes v1.9.7
etcd v3.3.4
docker 18.03.0-ce
calico/node:v3.0.6
calico/cni:v2.0.5
calico/kube-controllers:v2.0.4
centos 7.3+
集群规划和基础参数设定：
1.高可用集群所需节点配置如下：
部署节点×1：运行ansible脚本的节点
etcd节点×3：注意etcd集群节点必须是1，3，5，7……奇数个节点
master节点×1：运行集群主要组件
node节点×3：真正应用部署的节点，根据需要增加机器配置和节点数
2.在部署节点准备ansible：使用ansible的docker环境启动：
1° 下载内部源配置脚本并运行脚本：
wget http://download2.yunwei.edu/shell/yum-repo.sh
bash yum-repo.sh
2°下载并安装docker：
wget http://download2.yunwei.edu/shell/docker.tar.gz 解压后切换到docker目录下 运行docker.sh脚本 查看docker服务是否启动
docker image
#/bin/bash tar zxvf docker-app.tar.gz -C /usr/local/bin/ mkdir -p /etc/docker mkdir -p /etc/docker/certs.d/reg.yunwei.edu cp ca.crt /etc/docker/certs.d/reg.yunwei.edu/ echo "172.16.254.20 reg.yunwei.edu">>/etc/hosts cat <<EOF>/etc/docker/daemon.json { "registry-mirrors": ["http://cc83932c.m.daocloud.io"], "max-concurrent-downloads": 10, "log-driver": "json-file", "log-level": "warn", "log-opts": { "max-size": "10m", "max-file": "3" } } EOF cat <<EOF>/etc/systemd/system/docker.service [Unit] Description=Docker Application Container Engine Documentation=http://docs.docker.io [Service] Environment="PATH=/usr/local/bin:/bin:/sbin:/usr/bin:/usr/sbin" ExecStart=/usr/local/bin/dockerd ExecStartPost=/sbin/iptables -I FORWARD -s 0.0.0.0/0 -j ACCEPT ExecReload=/bin/kill -s HUP $MAINPID Restart=on-failure RestartSec=5 LimitNOFILE=infinity LimitNPROC=infinity LimitCORE=infinity Delegate=yes KillMode=process [Install] WantedBy=multi-user.target EOF systemctl daemon-reload && systemctl enable docker.service && systemctl start docker.service
3°下载并运行docker版的ansible：
docker pull reg.yunwei.edu/learn/ansible:alpine3
docker run -itd -v /etc/ansible:/etc/ansible -v /etc/kubernetes/:/etc/kubernetes/ -v /root/.kube:/root/.kube -v /usr/local/bin/:/usr/local/bin/ 1acb4fd5df5b /bin/sh
4° 配置每台机器之间主机名解析
5° 进入ansible容器，设置免密钥登录
测试各节点是否正常
ansible all -m ping
3.在部署节点上传ansible工作文件：
wget http://download2.yunwei.edu/shell/kubernetes.tar.gz
解压后：
bash harbor-offline-installer-v1.4.0.tgz k8s197.tar.gz scope.yaml
ca.tar.gz image.tar.gz kube-yunwei-197.tar.gz sock-shop
解压kube-yunwei-197.tar.gz 并将kube-yunwei-197下的所有文件移动到ansible下，删除kube-yunwei-197
[root@localhost kubernetes]# tar xf kube-yunwei-197.tar.gz [root@localhost kubernetes]# ls bash harbor-offline-installer-v1.4.0.tgz k8s197.tar.gz kube-yunwei-197.tar.gz sock-shop ca.tar.gz image.tar.gz kube-yunwei-197 scope.yaml [root@localhost kubernetes]# cd kube-yunwei-197 [root@localhost kube-yunwei-197]# ls 01.prepare.yml 03.docker.yml 05.kube-node.yml 99.clean.yml bin manifests tools 02.etcd.yml 04.kube-master.yml 06.network.yml ansible.cfg example roles [root@localhost kube-yunwei-197]# mv * /etc/ansible/
解压 k8s197.tar.gz 并将bin下的所有文件移动到ansible下的bin目录下，删除kubernetes下的bin
[root@localhost kubernetes]# tar xf k8s197.tar.gz [root@localhost kubernetes]# ls bash ca.tar.gz image k8s197.tar.gz kube-yunwei-197.tar.gz sock-shop bin harbor-offline-installer-v1.4.0.tgz image.tar.gz kube-yunwei-197 scope.yaml [root@localhost kubernetes]# cd bin [root@localhost bin]# ls bridge docker dockerd etcdctl kubectl portmap calicoctl docker-compose docker-init flannel kubelet cfssl docker-containerd docker-proxy host-local kube-proxy cfssl-certinfo docker-containerd-ctr docker-runc kube-apiserver kube-scheduler cfssljson docker-containerd-shim etcd kube-controller-manager loopback [root@localhost bin]# mv * /etc/ansible/bin/ [root@localhost bin]# ls [root@localhost bin]# cd /etc/ansible/bin/ [root@localhost bin]# ls bridge docker dockerd etcdctl kubectl portmap calicoctl docker-compose docker-init flannel kubelet VERSION.md cfssl docker-containerd docker-proxy host-local kube-proxy cfssl-certinfo docker-containerd-ctr docker-runc kube-apiserver kube-scheduler cfssljson docker-containerd-shim etcd kube-controller-manager loopback
切换到example目录下，将 hosts.s-master.example 文件复制到ansible目录下，并改名为hosts
[root@localhost kubernetes]# cd /etc/ansible/ [root@localhost ansible]# ls 01.prepare.yml 03.docker.yml 05.kube-node.yml 99.clean.yml bin manifests tools 02.etcd.yml 04.kube-master.yml 06.network.yml ansible.cfg example roles [root@localhost ansible]# cd example/ [root@localhost example]# ls hosts.s-master.example [root@localhost example]# cp hosts.s-master.example ../hosts [root@localhost example]# cd .. [root@localhost ansible]# ls 01.prepare.yml 03.docker.yml 05.kube-node.yml 99.clean.yml bin hosts roles 02.etcd.yml 04.kube-master.yml 06.network.yml ansible.cfg example manifests tools [root@localhost ansible]# vim hosts
# 在deploy节点生成CA相关证书，以及kubedns.yaml配置文件 - hosts: deploy roles: - deploy # 集群节点的公共配置任务 - hosts: - kube-master - kube-node - deploy - etcd - lb roles: - prepare # [可选]多master部署时的负载均衡配置 - hosts: lb roles: - lb
- hosts: etcd
roles:
- etcd
- hosts: - kube-master - kube-node roles: - docker
- hosts: kube-master roles: - kube-master - kube-node # 禁止业务 pod调度到 master节点 tasks: - name: 禁止业务 pod调度到 master节点 shell: "{{ bin_dir }}/kubectl cordon {{ NODE_IP }} " when: DEPLOY_MODE != "allinone" ignore_errors: true
- hosts: kube-node
roles:
- kube-node
# 集群网络插件部署，只能选择一种安装 - hosts: - kube-master - kube-node roles: - { role: calico, when: "CLUSTER_NETWORK == 'calico'" } - { role: flannel, when: "CLUSTER_NETWORK == 'flannel'" }
编辑hosts文件
# 部署节点：运行ansible 脚本的节点 [deploy] 192.168.42.30 # etcd集群请提供如下NODE_NAME、NODE_IP变量,请注意etcd集群必须是1,3,5,7...奇数个节点 [etcd] 192.168.42.121 NODE_NAME=etcd1 NODE_IP="192.168.42.121" 192.168.42.122 NODE_NAME=etcd2 NODE_IP="192.168.42.122" 192.168.42.172 NODE_NAME=etcd3 NODE_IP="192.168.42.172" [kube-master] 192.168.42.121 NODE_IP="192.168.42.121" [kube-node] 192.168.42.121 NODE_IP="192.168.42.121" 192.168.42.122 NODE_IP="192.168.42,122" 192.168.42.172 NODE_IP="192.168.42.172" [all:vars] # ---------集群主要参数--------------- #集群部署模式：allinone, single-master, multi-master DEPLOY_MODE=single-master #集群 MASTER IP MASTER_IP="192.168.42.121" #集群 APISERVER KUBE_APISERVER="https://192.168.42.121:6443" #TLS Bootstrapping 使用的 Token，使用 head -c 16 /dev/urandom | od -An -t x | tr -d ' ' 生成 BOOTSTRAP_TOKEN="d18f94b5fa585c7123f56803d925d2e7" # 集群网络插件，目前支持calico和flannel CLUSTER_NETWORK="flannel" # 部分calico相关配置，更全配置可以去roles/calico/templates/calico.yaml.j2自定义 # 设置 CALICO_IPV4POOL_IPIP=“off”,可以提高网络性能，条件限制详见 05.安装calico网络组件.md CALICO_IPV4POOL_IPIP="always" # 设置 calico-node使用的host IP，bgp邻居通过该地址建立，可手动指定端口"interface=eth0"或使用如下自动发现 IP_AUTODETECTION_METHOD="can-reach=223.5.5.5" # 部分flannel配置，详见roles/flannel/templates/kube-flannel.yaml.j2 FLANNEL_BACKEND="vxlan" # 服务网段 (Service CIDR），部署前路由不可达，部署后集群内使用 IP:Port 可达 SERVICE_CIDR="10.68.0.0/16" # POD 网段 (Cluster CIDR），部署前路由不可达，**部署后**路由可达 CLUSTER_CIDR="172.20.0.0/16" # 服务端口范围 (NodePort Range) NODE_PORT_RANGE="20000-40000" # kubernetes 服务 IP (预分配，一般是 SERVICE_CIDR 中第一个IP) CLUSTER_KUBERNETES_SVC_IP="10.68.0.1" # 集群 DNS 服务 IP (从 SERVICE_CIDR 中预分配) CLUSTER_DNS_SVC_IP="10.68.0.2" # 集群 DNS 域名 CLUSTER_DNS_DOMAIN="cluster.local." # etcd 集群间通信的IP和端口, **根据实际 etcd 集群成员设置** ETCD_NODES="etcd1=https://192.168.42.121:2380,etcd2=https://192.168.42.122:2380,etcd3=https://192.168.42.172:2380" # etcd 集群服务地址列表, **根据实际 etcd 集群成员设置** ETCD_ENDPOINTS="https://192.168.42.121:2379,https://192.168.42.122:2379,https://192.168.42.172:2379" # 集群basic auth 使用的用户名和密码 BASIC_AUTH_USER="admin" BASIC_AUTH_PASS="test1234" # ---------附加参数-------------------- #默认二进制文件目录 bin_dir="/usr/local/bin" #证书目录 ca_dir="/etc/kubernetes/ssl" #部署目录，即 ansible 工作目录 base_dir="/etc/ansible"
二、部署kubernetes过程
进入容器，查看ansible目录下是否有文件，并且查看能否ping通其他节点
[root@localhost ansible]# docker exec -it 0918862b8730 /bin/sh / # cd /etc/ansible/ /etc/ansible # ls 01.prepare.yml 06.network.yml hosts 02.etcd.yml 99.clean.yml manifests 03.docker.yml ansible.cfg roles 04.kube-master.yml bin tools 05.kube-node.yml example /etc/ansible # ansible all -m ping 192.168.42.122 | SUCCESS => { "changed": false, "ping": "pong" } 192.168.42.172 | SUCCESS => { "changed": false, "ping": "pong" } 192.168.42.121 | SUCCESS => { "changed": false, "ping": "pong" } 192.168.42.30 | SUCCESS => { "changed": false, "ping": "pong" }
ansible-playbook 01.prepare.yml ansible-playbook 02.etcd.yml ansible-playbook 03.docker.yml ansible-playbook 04.kube-master.yml ansible-playbook 05.kube-node.yml
在执行06.network.yml之前要确保其他节点有镜像，所以解压image.tar.gz
[root@cicd kubernetes]# ls bash ca.tar.gz image k8s197.tar.gz kube-yunwei-197 scope.yaml bin harbor-offline-installer-v1.4.0.tgz image.tar.gz kubernetes.tar.gz kube-yunwei-197.tar.gz sock-shop [root@cicd kubernetes]# cd image [root@cicd image]# ls calico-cni-v2.0.5.tar coredns-1.0.6.tar.gz influxdb-v1.3.3.tar calico-kube-controllers-v2.0.4.tar grafana-v4.4.3.tar kubernetes-dashboard-amd64-v1.8.3.tar.gz calico-node-v3.0.6.tar heapster-v1.5.1.tar pause-amd64-3.1.tar [root@cicd image]# scp ./* node1:/root/image
[root@cicd image]# scp ./* node2:/root/image
[root@cicd image]# scp ./* node3:/root/image
在node节点：
[root@node1 image]# for i in `ls`;do docker load -i $i;done
部署节点：
ansible-playbook 06.network.yml
/etc/ansible # ls 01.prepare.yml 03.docker.yml 06.network.yml bin manifests 02.etcd.retry 04.kube-master.yml 99.clean.yml example roles 02.etcd.yml 05.kube-node.yml ansible.cfg hosts tools /etc/ansible # cd manifests/ /etc/ansible/manifests # ls coredns dashboard efk heapster ingress kubedns /etc/ansible/manifests # cd coredns/ /etc/ansible/manifests/coredns # ls coredns.yaml /etc/ansible/manifests/coredns # kubectl create -f . serviceaccount "coredns" created clusterrole "system:coredns" created clusterrolebinding "system:coredns" created configmap "coredns" created deployment "coredns" created service "coredns" created
/etc/ansible/manifests # ls
coredns dashboard efk heapster ingress kubedns
/etc/ansible/manifests # cd dashboard/
/etc/ansible/manifests/dashboard # ls
1.6.3 kubernetes-dashboard.yaml ui-read-rbac.yaml
admin-user-sa-rbac.yaml ui-admin-rbac.yaml
/etc/ansible/manifests/dashboard # kubectl create -f .
serviceaccount "admin-user" created
clusterrolebinding "admin-user" created
secret "kubernetes-dashboard-certs" created
serviceaccount "kubernetes-dashboard" created
role "kubernetes-dashboard-minimal" created
rolebinding "kubernetes-dashboard-minimal" created
deployment "kubernetes-dashboard" created
service "kubernetes-dashboard" created
clusterrole "ui-admin" created
rolebinding "ui-admin-binding" created
clusterrole "ui-read" created
rolebinding "ui-read-binding" created
[root@cicd ansible]# kubectl get pod -n kube-system -o wide NAME READY STATUS RESTARTS AGE IP NODE coredns-6ff7588dc6-l8q4h 1/1 Running 0 7m 172.20.0.2 192.168.42.122 coredns-6ff7588dc6-x2jq5 1/1 Running 0 7m 172.20.1.2 192.168.42.172 kube-flannel-ds-c688h 1/1 Running 0 14m 192.168.42.172 192.168.42.172 kube-flannel-ds-d4p4j 0/1 Running 0 14m 192.168.42.122 192.168.42.122 kube-flannel-ds-f8gp2 1/1 Running 0 14m 192.168.42.121 192.168.42.121 kubernetes-dashboard-545b66db97-z9nr4 1/1 Running 0 1m 172.20.1.3 192.168.42.172
[root@cicd ansible]# kubectl cluster-info Kubernetes master is running at https://192.168.42.121:6443 CoreDNS is running at https://192.168.42.121:6443/api/v1/namespaces/kube-system/services/coredns:dns/proxy kubernetes-dashboard is running at https://192.168.42.121:6443/api/v1/namespaces/kube-system/services/https:kubernetes-dashboard:/proxy To further debug and diagnose cluster problems, use 'kubectl cluster-info dump'.
密钥获取：
kubectl -n kube-system describe secret $(kubectl -n kube-system get secret|grep admin-user|awk '{print $1}')