When I first started my Kubernetes trip, I was looking for a way to set up local deployment environment. Many people often use minikube or microk8s, both for new users to operate in a single-node cluster environment. But when I already know the basics, both of which is clearly not enough, I need to be able to run further to find the local multi-node cluster, the production environment more similar platform. To this end, I have access to many resources, and finally I found Rancher Server. Next, I'll show you how I set up my local K8S multi-node cluster.
Prepare master node and worker nodes of a virtual machine
The figure shows the architecture of a cluster, a worker master node and three nodes. To try to use ingress controller such as load balancing, session remains, host affinity and other cross-node function, which is the minimum configuration.
Rancher Server installation is not complicated, it is packaged as a docker image, and can be run as a container. The basic configuration of a node is running docker CE daemon on Linux virtual machines. In this case, I chose to use Ubuntu 18.04 LTS (Please refer to the specific requirements of the node Rancher official document: https: //rancher.com/docs/rancher/v2.x/en/installation/requirements/). When I finished the first VM settings, I simply copy it to the other three. If you and I are the same operation, refer to below two tips:
Install the specified version Docker: https: //docs.docker.com/install/linux/docker-ce/ubuntu/
After the Clone virtual machine, change the Ubuntu hostname: https: //linuxize.com/post/how-to-change-hostname-on-ubuntu-18-04/
Eventually, the four nodes on my computer (i5,24G RAM) starts, and assign the following resources.
Master节点（2 core、4G RAM、Ubuntu 18.04 + Docker CE 18.09) x 1
Worker节点（2 core, 3G RAM, Ubuntu 18.04 + Docker CE 18.09) x 3
在master节点上启动Rancher server
sudo docker run --restart=unless-stopped \
-p 81:80 -p 444:443 rancher/rancher以上命令可以启动Rancher server容器，并且将其运行在master节点上。默认状态下，nginx ingress controller会嵌入到worker节点中，并且绑定端口80和443。因此我将Rancger server发布到端口81和444或其他端口，以避免端口冲突。
完成Rancher server初始设置
首先，使用master节点IP地址和端口444启动Rancher server控制台，它会要求设置管理员密码。
接着，确认worker节点到达Rancher server的URL。这里为了方便我直接使用master节点IP地址作为URL。完成了初始设置后，Rancher server就可以添加新集群了。
创建一个新的K8S集群和master节点
因为我想要在本地虚拟机上运行K8S集群而不是在云端，所以选择“自定义”的选项（即上方说明为“ From my own existing nodes”），然后新集群的云提供商选项选为“None”。
在ubuntu虚拟机上复制并运行docker命令来启动master节点。一个master节点至少要有etcd和control，如果你打算创建一个单节点集群，需要选择所有3个角色并更改命令。
运行docker命令之后，新节点将显示在Rancher Server控制台上，配置这一节点需要花费一些时间，一旦配置完成，状态将变为active。
创建worker节点
对于worker节点，我们仅需要在节点选项选择“Worker”角色，然后在3个worker节点的Ubuntu虚拟机上复制并运行docker命令。
最后，在我的电脑上多节点集群已经准备就绪。
安装kubectl工具来管理新的K8S集群
新集群的Kubernetes版本时v1.14.6，你可以在上面的截图内看到。为了更好地将kubectl工具版本与集群匹配，在master节点上运行以下命令，以安装特定版本：
curl -LO https://storage.googleapis.com/kubernetes-release/release/v1.14.6/bin/linux/amd64/kubectl
chmod +x ./kubectl
sudo mv ./kubectl /usr/local/bin/kubectlKubectl工具需要一个kubeconfig文件与集群连接，新集群地kubeconfig文件可以在Rancher Server的控制台中找到。
复制以上kubeconfig文件，并保存为~/.kube/config文件。随后kubectl就能够获取集群信息。
有关kubectl的安装和配置可以参考以下连接：
安装和设置kubectl：https://kubernetes.io/docs/tasks/tools/install-kubectl/
配置kubectl以访问多集群：https://kubernetes.io/docs/tasks/access-application-cluster/configure-access-multiple-clusters/
原文链接：
https://medium.com/@kwonghung.yip/setup-local-kubernetes-multi-node-cluster-with-rancher-server-fdb7a0669b5c