。K8S是服务（容器）的管理工具，它支持集群架构。集群的nodes们要先在infrastructure上建起来。
使用kubeadm 在infrastructure节点上创建K8S集群
step 1. 在所有nodes上安装 集群管理工具 kubeadm
https://kubernetes.io/docs/setup/production-environment/tools/kubeadm/install-kubeadm/
A. 添加google cloud repo
[root@afdev2 ~]# cat <
> [kubernetes]
> name=Kubernetes
> baseurl=https://packages.cloud.google.com/yum/repos/kubernetes-el7-x86_64
> enabled=1
> gpgcheck=1
> repo_gpgcheck=1
> gpgkey=https://packages.cloud.google.com/yum/doc/yum-key.gpg https://packages.cloud.google.com/yum/doc/rpm-package-key.gpg
> EOF
B. 设置SELinux
[root@afdev2 ~]# setenforce 0
setenforce: SELinux is disabled
[root@afdev2 ~]# sed -i 's/^SELINUX=enforcing$/SELINUX=permissive/' /etc/selinux/config
C. 解决依赖关系：1/2是kuberctl依赖，3/4是conntrack依赖
1. install socat --> TCP port forwarder https://centos.pkgs.org/7/centos-x86_64/socat-1.7.3.2-2.el7.x86_64.rpm.html
[root@afdev2 socat]# rpm -ivh socat-1.7.3.2-2.el7.x86_64.rpm
2. install conntrack --> Netfilter's connection tracking userspace tools https://centos.pkgs.org/7/centos-x86_64/conntrack-tools-1.4.4-4.el7.x86_64.rpm.html
3. install libnetfilter_cthelper --> http://rpm.pbone.net/index.php3/stat/4/idpl/27123498/dir/redhat_el_7/com/libnetfilter_cthelper-1.0.0-4.el7.x86_64.rpm.html
4. install libnetfilter_cttimeout and libnetfilter_queue --> [root@afdev2 conntrack]# yum install libnetfilter_cttimeout libnetfilter_queue
D. 安装kubeadm *
[root@afdev2 k8s_sw]# yum install -y kubelet kubeadm kubectl --disableexcludes=kubernetes
Installing:
kubeadm x86_64 1.15.0-0 kubernetes 8.9 M
kubectl x86_64 1.15.0-0 kubernetes 9.5 M
kubelet x86_64 1.15.0-0 kubernetes 22 M
Installing for dependencies:
cri-tools x86_64 1.13.0-0 kubernetes 5.1 M
kubernetes-cni x86_64 0.7.5-0 kubernetes 10 M
E. 装载br_netfilter
# modprobe br_netfilter
F. 确保K8S trafiic 不被iptables bypass
[root@afdev2 ~]# cat <
> net.bridge.bridge-nf-call-ip6tables = 1
> net.bridge.bridge-nf-call-iptables = 1
> EOF
[root@afdev2 ~]# sysctl --system
G. Docker 安装。因为kube-apiserver等等这些kube组件都将以容器形式运行
# yum install docker
H. 确保 server hostame解析ok（/etc/hosts），CPU不少于2个， disable swap （swapoff -a）
I. start/enable docker和kubelet service
[root@afdev2 ~]# systemctl enable --now docker
Created symlink from /etc/systemd/system/multi-user.target.wants/docker.service to /usr/lib/systemd/system/docker.service.
[root@afdev2 ~]# systemctl enable --now kubelet
Created symlink from /etc/systemd/system/multi-user.target.wants/kubelet.service to /usr/lib/systemd/system/kubelet.service.
[root@afdev2 ~]#
step 2. 在选做master的node上初始化集群
# kuberadm init --> output的最后一行，token。所有工作节点加入集群的认证令牌
# mkdir -p $HOME/.kube
# cp -i /etc/kubernetes/admin.conf $HOME/.kube/config
# chown $(id -u):$(id -g) $HOME/.kube/config
step 3. 让选做工作节点（后面叫nodes）加入集群
# kuberadm join
嘻嘻 log在<2019-07-18 kubeadm init/join output>
step 4. 为了配置master和node、和以后的pods要求的网络，需要给master安装网络插件
现在node的状态都是NOT READY， 正是因为少这个：
[root@afdev2 ~]# kubectl getnodes
NAME STATUS ROLES AGE VERSION
afdev1 NotReady 16h v1.15.1
afdev2 NotReady master 17h v1.15.0
安装网路插件 Weave Net
[root@afdev2 ~]# kubectl version
Client Version: version.Info{Major:"1", Minor:"15", GitVersion:"v1.15.0", GitCommit:"e8462b5b5dc2584fdcd18e6bcfe9f1e4d970a529", GitTreeState:"clean", BuildDate:"2019-06-19T16:40:16Z", GoVersion:"go1.12.5", Compiler:"gc", Platform:"linux/amd64"}
Server Version: version.Info{Major:"1", Minor:"15", GitVersion:"v1.15.1", GitCommit:"4485c6f18cee9a5d3c3b4e523bd27972b1b53892", GitTreeState:"clean", BuildDate:"2019-07-18T09:09:21Z", GoVersion:"go1.12.5", Compiler:"gc", Platform:"linux/amd64"}
[root@afdev2 ~]# export kubever=$(kubectl version | base64 | tr -d '\n')
[root@afdev2 ~]# kubectl apply -f https://cloud.weave.works/k8s/net?k8s-version=$kubever
serviceaccount/weave-net created
clusterrole.rbac.authorization.k8s.io/weave-net created
clusterrolebinding.rbac.authorization.k8s.io/weave-net created
role.rbac.authorization.k8s.io/weave-net created
rolebinding.rbac.authorization.k8s.io/weave-net created
daemonset.extensions/weave-net created
所有节点都READY了 :)
[root@afdev2 ~]# kubectl get nodes
NAME STATUS ROLES AGE VERSION
afdev1 Ready
afdev2 Ready master 17h v1.15.0
[root@afdev2 ~]#
* 如果想选择其他的网络插件，可以从这里下载和找到安装方法：https://kubernetes.io/docs/concepts/cluster-administration/addons/
K8S的组件是这样的：
master上：
etcd --> port 2379。1个daemon，many socket ，全来自api。 (Openshift : port一样；1个Daemon，除了many socket 来自api，还有1个来自controller-manager)
* etcd是一种键值分布是存储。etcdctl是管理工具。
[root@afdev2 hack]# docker exec -it k8s_etcd_etcd-afdev2_kube-system_243c0cac85ba38f13f7517899628e9a1_0 /bin/sh
/ # export ETCDCTL_API=3
/ # etcdctl member list
Error: context deadline exceeded
identify secure client using this TLS certificate/key file :
[root@afdev2 ~]# docker inspect k8s_etcd_etcd-afdev2_kube-system_243c0cac85ba38f13f7517899628e9a1_0 | egrep -i "(peer-cert|peer-key)"
"--peer-cert-file=/etc/kubernetes/pki/etcd/peer.crt",
"--peer-key-file=/etc/kubernetes/pki/etcd/peer.key",
"--peer-cert-file=/etc/kubernetes/pki/etcd/peer.crt",
"--peer-key-file=/etc/kubernetes/pki/etcd/peer.key",
[root@afdev2 ~]#
/ # ls -l /etc/kubernetes/pki/etcd/
total 32
-rw-r--r-- 1 root root 1017 Jul 18 18:46 ca.crt
-rw------- 1 root root 1679 Jul 18 18:46 ca.key
-rw-r--r-- 1 root root 1094 Jul 18 18:46 healthcheck-client.crt
-rw------- 1 root root 1679 Jul 18 18:46 healthcheck-client.key
-rw-r--r-- 1 root root 1127 Jul 18 18:46 peer.crt
-rw------- 1 root root 1679 Jul 18 18:46 peer.key
-rw-r--r-- 1 root root 1127 Jul 18 18:46 server.crt
-rw------- 1 root root 1679 Jul 18 18:46 server.key
/ #
/ # ETCDCTL_API=3 etcdctl --cert /etc/kubernetes/pki/etcd/server.crt --key /etc/kubernetes/pki/etcd/server.key get --prefix / --keys-only
Error: context deadline exceeded
/ # ETCDCTL_API=3 etcdctl --cacert="/etc/kubernetes/pki/etcd/ca.crt" --cert="/etc/kubernetes/pki/etcd/peer.crt" --key="/etc/kubernetes/pki/etcd/peer.key" get --prefix / --keys-only --> :) 列出了所有键名
/registry/apiregistration.k8s.io/apiservices/v1.
/registry/apiregistration.k8s.io/apiservices/v1.apps
/registry/apiregistration.k8s.io/apiservices/v1.authentication.k8s.io
/registry/apiregistration.k8s.io/apiservices/v1.authorization.k8s.io
/registry/apiregistration.k8s.io/apiservices/v1.autoscaling
/registry/apiregistration.k8s.io/apiservices/v1.batch
让我看一个键值：
/ # ETCDCTL_API=3 etcdctl --cacert="/etc/kubernetes/pki/etcd/ca.crt" --cert="/etc/kubernetes/pki/etcd/peer.crt" --key="/etc/kubernetes/pki/etcd/peer.key" get /registry/
apiregistration.k8s.io/apiservices/v1.storage.k8s.io
/registry/apiregistration.k8s.io/apiservices/v1.storage.k8s.io
{"kind":"APIService","apiVersion":"apiregistration.k8s.io/v1beta1","metadata":{"name":"v1.storage.k8s.io","uid":"d1fed909-48ef-4654-91c7-533d58ceaf10","creationTimestamp":"2019-07-18T18:46:38Z","labels":{"kube-aggregator.kubernetes.io/automanaged":"onstart"}},"spec":{"service":null,"group":"storage.k8s.io","version":"v1","groupPriorityMinimum":16800,"versionPriority":15},"status":{"conditions":[{"type":"Available","status":"True","lastTransitionTime":"2019-07-18T18:46:38Z","reason":"Local","message":"Local APIServices are always available"}]}}
scheduler --> port NA。不是daemon 没有监听端口。 1个 进程，1个socket
api --> port 6443/sun-sr-https。 1个Daemon，socket= 其他组建（除了etcd）的client connection socket数 + nodes上的kubelet（n）和proxy（1） 。作为etcd的客户端， 向etcd daemon有many socket (Openshift : port 8443/pcsync-https。）
controller-manager --> port NA。不是daemon 没有监听端口。一个进程里，n个socket
kubelet --> port NA。不是daemon 没有监听端口。一个进程里，n个socket
proxy --> port NA。不是daemon 没有监听端口。一个进程里，1个socket
nodes上：
proxy --> port NA。不是daemon 没有监听端口。一个进程里，1个socket。连向master
kubelet --> port NA。不是daemon 没有监听端口。一个进程里，n个socket。连向master
K8S的工具又是这样的：
master上：
kubeadm --> bootstrap
kubectl --> talk to nodes
nodes上：
kubelet --> starting pods and container