本文介绍在centos7.4使用kubeadm快速离线安装kubernetes 1.10。
采用单master,单node(可以多node),占用资源较少,方便在笔记本或学习环境快速部署,不适用于生产环境。
所需文件百度盘连接
链接:https://pan.baidu.com/s/179xcJWWeQpAX7ctxABeWyw 密码:ep50
主机名 | IP | 配置 |
---|---|---|
master1 | 192.168.145.220 | 2G内存,20G存储,4核CPU |
node1 | 192.168.145.221 | 2G内存,20G存储,4核CPU |
准备不低于2台虚机。 1台 master,其余的做node
OS: Centos7.4 mini install。 最小化安装。配置节点IP
主机名 | IP | 配置 |
---|---|---|
master1 | 192.168.145.220 | 2G内存,20G存储,4核CPU |
node1 | 192.168.145.221 | 2G内存,20G存储,4核CPU |
分别设置主机名为master1 node1 ... 时区
timedatectl set-timezone Asia/Shanghai #都要执行
hostnamectl set-hostname master1 #master1执行
hostnamectl set-hostname node1 #node1执行
在所有节点/etc/hosts中添加解析,master1,node1
127.0.0.1 localhost localhost.localdomain localhost4 localhost4.localdomain4
::1 localhost localhost.localdomain localhost6 localhost6.localdomain6
192.168.1.181 matser1
192.168.1.182 node1
关闭所有节点的seliux以及firewalld
sed -i 's/SELINUX=enforcing/SELINUX=disabled/g' /etc/selinux/config
setenforce 0
systemctl disable firewalld
systemctl stop firewalld
使用文件docker-packages.tar,每个节点都要安装。【文件来源是在百度云盘】
tar -xvf docker-packages.tar
cd docker-packages
rpm -Uvh * 或者 yum install local *.rpm 进行安装
docker version #安装完成查看版本
启动docker,并设置为开机自启
systemctl start docker && systemctl enable docker
输入docker info,==记录Cgroup Driver==
Cgroup Driver: cgroupfs
docker和kubelet的cgroup driver需要一致,如果docker不是cgroupfs,则执行
cat << EOF > /etc/docker/daemon.json
{
"exec-opts": ["native.cgroupdriver=cgroupfs"]
}
EOF
systemctl daemon-reload && systemctl restart docker
使用文件kube-packages-1.10.1.tar,每个节点都要安装
kubeadm是集群部署工具
kubectl是集群管理工具,通过command来管理集群
kubelet的k8s集群每个节点的docker管理服务
tar -xvf kube-packages-1.10.1.tar
cd kube-packages-1.10.1
rpm -Uvh * 或者 yum install local *.rpm 进行安装
在所有kubernetes节点上设置kubelet使用cgroupfs,与dockerd保持一致,否则kubelet会启动报错
默认kubelet使用的cgroup-driver=systemd,改为cgroup-driver=cgroupfs
sed -i "s/cgroup-driver=systemd/cgroup-driver=cgroupfs/g" /etc/systemd/system/kubelet.service.d/10-kubeadm.conf
重设kubelet服务,并重启kubelet服务
systemctl daemon-reload && systemctl restart kubelet
关闭swap,及修改iptables,不然后面kubeadm会报错
swapoff -a
vi /etc/fstab #swap一行注释
cat < /etc/sysctl.d/k8s.conf
net.bridge.bridge-nf-call-ip6tables = 1
net.bridge.bridge-nf-call-iptables = 1
EOF
sysctl --system
使用文件k8s-images-1.10.tar.gz,每个节点都要执行
节点较少,就不搭建镜像仓库服务了,后续要用的应用镜像,每个节点都要导入
docker load -i k8s-images-1.10.tar.gz
一共11个镜像,分别是
k8s.gcr.io/etcd-amd64:3.1.12
k8s.gcr.io/kube-apiserver-amd64:v1.10.1
k8s.gcr.io/kube-controller-manager-amd64:v1.10.1
k8s.gcr.io/kube-proxy-amd64:v1.10.1
k8s.gcr.io/kube-scheduler-amd64:v1.10.1
k8s.gcr.io/pause-amd64:3.1
k8s.gcr.io/k8s-dns-kube-dns-amd64:1.14.8
k8s.gcr.io/k8s-dns-dnsmasq-nanny-amd64:1.14.8
k8s.gcr.io/k8s-dns-sidecar-amd64:1.14.8
k8s.gcr.io/kubernetes-dashboard-amd64:v1.8.3
quay.io/coreos/flannel:v0.9.1-amd64
只在master执行。此处选用最简单快捷的部署方案。etcd、api、controller-manager、 scheduler服务都会以容器的方式运行在master。etcd 为单点,不带证书。etcd的数据会挂载到master节点/var/lib/etcd
init部署是支持etcd 集群和证书模式的,配置方法见我1.9的文档,此处略过。
init命令注意要指定版本,和pod范围
kubeadm init --kubernetes-version=v1.10.1 --pod-network-cidr=10.244.0.0/16
Your Kubernetes master has initialized successfully!
To start using your cluster, you need to run the following as a regular user:
mkdir -p $HOME/.kube
sudo cp -i /etc/kubernetes/admin.conf $HOME/.kube/config
sudo chown $(id -u):$(id -g) $HOME/.kube/config
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 node
as root:
kubeadm join 192.168.145.226:6443 --token 1tlh4c.6g1ss2tu1aq9p3pe --discovery-token-ca-cert-hash sha256:2b0ba1ecf33866aff93b51ce11ebe44ef3e0fa13703edc5a310f55bd3db59060
记下join的命令,后续node节点加入的时候要用到
执行提示的命令,保存kubeconfig
mkdir -p $HOME/.kube
sudo cp -i /etc/kubernetes/admin.conf $HOME/.kube/config
sudo chown $(id -u):$(id -g) $HOME/.kube/config
此时执行kubectl get node 已经可以看到master节点,notready是因为还未部署网络插件
[root@master1 kubernetes1.10]# kubectl get node
NAME STATUS ROLES AGE VERSION
master1 NotReady master 3m v1.10.1
查看所有的pod,kubectl get pod --all-namespaces
kubedns也依赖于容器网络,此时pending是正常的
[root@master1 kubernetes1.10]# kubectl get pod --all-namespaces
NAMESPACE NAME READY STATUS RESTARTS AGE
kube-system etcd-master1 1/1 Running 0 3m
kube-system kube-apiserver-master1 1/1 Running 0 3m
kube-system kube-controller-manager-master1 1/1 Running 0 3m
kube-system kube-dns-86f4d74b45-5nrb5 0/3 Pending 0 4m
kube-system kube-proxy-ktxmb 1/1 Running 0 4m
kube-system kube-scheduler-master1 1/1 Running 0 3m
配置KUBECONFIG变量
echo "export KUBECONFIG=/etc/kubernetes/admin.conf" >> /etc/profile
source /etc/profile
echo $KUBECONFIG #应该返回/etc/kubernetes/admin.conf
k8s支持多种网络方案,flannel,calico,openvswitch
此处选择flannel。 在熟悉了k8s部署后,可以尝试其他网络方案,我另外一篇1.9部署中有介绍flannel和calico的方案,以及切换时需要的动作。
kubectl apply -f kube-flannel.yml
网络就绪后,节点的状态会变为ready
[root@master1 kubernetes1.10]# kubectl get node
NAME STATUS ROLES AGE VERSION
master1 Ready master 18m v1.10.1
6.1 node节点加入集群
使用之前kubeadm init 生产的join命令,加入成功后,回到master节点查看是否成功
在node节点执行join命令
kubeadm join 192.168.145.226:6443 --token 1tlh4c.6g1ss2tu1aq9p3pe --discovery-token-ca-cert-hash sha256:2b0ba1ecf33866aff93b51ce11ebe44ef3e0fa13703edc5a310f55bd3db59060
回到master
[root@master1 kubernetes1.10]# kubectl get node
NAME STATUS ROLES AGE VERSION
master1 Ready master 31m v1.10.1
node1 Ready 44s v1.10.1
至此,集群已经部署完成。
6.2 如果出现x509这个报错
如果有报错才需要做这一步,不然不需要。
这是因为master节点缺少KUBECONFIG变量
[discovery] Failed to request cluster info, will try again: [Get https://192.168.1.181:6443/api/v1/namespaces/kube-public/configmaps/cluster-info: x509: certificate has expired or is not yet valid]
master节点执行
export KUBECONFIG=$HOME/.kube/config
node节点kubeadm reset 再join
kubeadm reset
kubeadm join xxx ...
6.3 如果忘了join命令,加入节点方法
若node已经成功加入,忽略这一步。
使用场景:忘了保存上面kubeadm init生产的join命令,可按照下面的方法加入node节点。
首先master节点获取token,如果token list内容为空,则kubeadm token create创建一个,记录下token数据
[root@master1 kubernetes1.10]# kubeadm token list
TOKEN TTL EXPIRES USAGES DESCRIPTION EXTRA GROUPS
wct45y.tq23fogetd7rp3ck 22h 2018-06-02T21:38:57+08:00 authentication,signing The default bootstrap token generated by 'kubeadm init'. system:bootstrappers:kubeadm:default-node-token
node节点执行如下,把token部分进行替换
kubeadm join 192.168.145.226:6443 --token wct45y.tq23fogetd7rp3ck 192.168.1.181:6443 --discovery-token-unsafe-skip-ca-verification
dashboard是官方的k8s 管理界面,可以查看应用信息及发布应用。dashboard的语言是根据浏览器的语言自己识别的
官方默认的dashboard为https方式,如果用chrome访问会拒绝。本次部署做了修改,方便使用,使用了http方式,用chrome访问正常。
一共需要导入3个yaml文件
kubectl apply -f kubernetes-dashboard-http.yaml
kubectl apply -f admin-role.yaml
kubectl apply -f kubernetes-dashboard-admin.rbac.yaml
[root@master1 kubernetes1.10]# kubectl apply -f kubernetes-dashboard-http.yaml
serviceaccount "kubernetes-dashboard" created
role.rbac.authorization.k8s.io "kubernetes-dashboard-minimal" created
rolebinding.rbac.authorization.k8s.io "kubernetes-dashboard-minimal" created
deployment.apps "kubernetes-dashboard" created
service "kubernetes-dashboard" created
[root@master1 kubernetes1.10]# kubectl apply -f admin-role.yaml
clusterrolebinding.rbac.authorization.k8s.io "kubernetes-dashboard" created
[root@master1 kubernetes1.10]# kubectl apply -f kubernetes-dashboard-admin.rbac.yaml
clusterrolebinding.rbac.authorization.k8s.io "dashboard-admin" created
创建完成后,通过 http://任意节点的IP:31000即可访问ui
这两部分后续会单独写如何部署。
9.1 kubectl 命令补全
root@master1:/# vim /etc/profilecd #添加下面这句,再source
source <(kubectl completion bash)
root@master1:/# source /etc/profile
9.2 master节点默认不可部署pod
执行如下,node-role.kubernetes.io/master 可以在 kubectl edit node master1中taint配置参数下查到
root@master1:/var/lib/kubelet# kubectl taint node master1 node-role.kubernetes.io/master-
node "master1" untainted
9.3 node节点pod无法启动/节点删除网络重置
node1之前反复添加过,添加之前需要清除下网络
root@master1:/var/lib/kubelet# kubectl get po -o wide
NAME READY STATUS RESTARTS AGE IP NODE
nginx-8586cf59-6zw9k 1/1 Running 0 9m 10.244.3.3 node2
nginx-8586cf59-jk5pc 0/1 ContainerCreating 0 9m node1
nginx-8586cf59-vm9h4 0/1 ContainerCreating 0 9m node1
nginx-8586cf59-zjb84 1/1 Running 0 9m 10.244.3.2 node2
root@node1:~# journalctl -u kubelet
failed: rpc error: code = Unknown desc = NetworkPlugin cni failed to set up pod "nginx-8586cf59-rm4sh_default" network: failed to set bridge addr: "cni0" already has an IP address different from 10.244.2.1/24
12252 cni.go:227] Error while adding to cni network: failed to set bridge addr: "cni0" already
重置kubernetes服务,重置网络。删除网络配置,link
kubeadm reset
systemctl stop kubelet
systemctl stop docker
rm -rf /var/lib/cni/
rm -rf /var/lib/kubelet/*
rm -rf /etc/cni/
ifconfig cni0 down
ifconfig flannel.1 down
ifconfig docker0 down
ip link delete cni0
ip link delete flannel.1
systemctl start docker
加入节点
systemctl start docker
kubeadm join 192.168.145.226:6443 --token 1tlh4c.6g1ss2tu1aq9p3pe --discovery-token-ca-cert-hash sha256:2b0ba1ecf33866aff93b51ce11ebe44ef3e0fa13703edc5a310f55bd3db59060
注意:要想在node节点执行kubectl get pod 或者其他的命令,需要再node上执行以下命令:
sudo cp /etc/kubernetes/admin.conf $HOME/
#如果node节点上没有admin.conf的文件,可以从主节点复制过来
#主节点执行,xxx从节点ip
scp /etc/kubernetes/admin.conf xxx:/root/
#从节点执行
sudo chown $(id -u):$(id -g) $HOME/admin.conf
export KUBECONFIG=$HOME/admin.conf
不然会出现以下错误:The connection to the server localhost:8080 was refused - did you specify the right host or port?(单独开终端也会出现)