On Centos7, RPMs are from Ceontos7 extras repository

1 master, 2 nodes

• On master:
  • kubernetes-master, kubernetes-client - 1.5.2-0.2.gitc55cf2b.el7
  • etc - 3.1.0-2.el7
  • flannel - 0.7.0-1.el7
• On nodes:
  • kubernetes-client, kubernetes-node - 1.5.2-0.2.gitc55cf2b.el7

• On master:
  • etcd
  • kube-apiserver
  • kube-controller-manager
  • kube-scheduler
  • flanneld
• On nodes:
  • kubelet
  • kube-proxy
  • flanneld

• /etc/kubernetes/config - for both master and node

KUBE_LOGTOSTDERR=“–logtostderr=true”

KUBE_LOG_LEVEL="--v=0"
KUBE_ALLOW_PRIV="--allow-privileged=false"
KUBE_MASTER="--master=http://node-1:8080"
* /etc/kubernetes/apiserver
* 

• master and node IPs
  • master: node-1(192.168.50.101/24)
  • nodes: node2(192.168.50.102/24), node-3(192.168.50.103/24)
  • kubernetes cluster IP range: 10.254.0.0/16
  • flannel: 172.30.0.0/16, subnetLen:24, type: vxlan

• etcd: 2379
• kube-apiserver: 8080
• kubelet: 10250

$ kubectl run hello-node --image=gcr.io/google-samples/node-hello:1.0 --port=8080

$ kubectl get deployment hello-node
NAME         DESIRED   CURRENT   UP-TO-DATE   AVAILABLE   AGE
hello-node   1         1         1            1           2d

$ kubectl get pods
NAME                          READY     STATUS    RESTARTS   AGE
hello-node-3526609615-5slx8   1/1       Running   0          2d

$ kubectl expose deployment hello-node --type=NodePort

$ kubectl get service hello-node
NAME         CLUSTER-IP      EXTERNAL-IP   PORT(S)          AGE
hello-node   10.254.65.168   <nodes>       8080:32260/TCP   2d

$ kubectl get pods --selector="run=hello-node" --output=wide
NAME                          READY     STATUS    RESTARTS   AGE       IP            NODE
hello-node-3526609615-5slx8   1/1       Running   0          2d        172.30.65.2   node-2

$ curl http://node-2:32260
Hello Kubernetes!