k8s cluster of
k8s deployment environment
https://kubernetes.io/docs/setup/independent/create-cluster-kubeadm/#prerequisites
01 Installation
About installation, remove the current K8S cluster services provided by the cloud service provider itself, a local cluster installation are mainly three kinds:
- Based on the binary installation
- Based on the official installation of the tool kubeadm
- Installation of third-party tools based on
1, based on component deployment yum directly related to the system-level daemon disadvantages: controller-manager failure, the need to manually restart;
Node network: 192.168.0.0/16
Pod network address: 10.244.0.0/16
Service Network Address: 10.96.0.0/12 // mask is a 12-bit
etcd deployed on the master node, is not independent;
2, using kubeadm deployed, the deployment of a core component k8s pod, but requires manual deployment and in the master kubelet and docker Node ( docker engine is running the container, kubelet pod is a core component of the operating container );
Docker run using kubeadm node deployment, each node needs as k8s underlayer, i.e., including the master needs to install docker, while each node needs to be installed, run kubelet, docker, and to ensure that both the up and running. The first node is initialized to master, the other node is initialized to node. The API Server component on the master, etcd, controller-manager, scheduler runs as a pod (pod are static, not k8s management, just run in the form of the pod), which are run on kubelet, to containerized run of form (k8s own components operate as a pod). Node kube-proxy will operate as Pod. Each node includes a master needs to run flannel pod (pod dynamic, self-management by the k8s),
k8s在github上的地址:https://github.com/kubernetes
kubeadm初始化并安装、生成k8s集群的过程
https://github.com/kubernetes/kubeadm/blob/master/docs/design/design_v1.10.md
kubeadm初始化、安装集群的步骤:
1、master、nodes:安装kubelet、kubeadm、docker
2、master:kubeadm init //在master运行kubeadm init,初始化集群
3、nodes:kubeadm join //在各nodes节点执行kubeadm join,将节点加入到集群中去
安装实践
master:IP:192.168.184.141 主机名:node1
node1:IP:192.168.184.142 主机名:node2
node2:IP:192.168.184.142 主机名:node3