Deploy Kubernetes in Docker

Deploy Kubernetes in Docker

Kubernetes is an open-source container management framework of Google. It provides a series of functions such as host, cluster management, container deployment, high availability, and auto scaling of Docker containers; kubernetes's design goals include putting the container cluster in the desired state at any time. Therefore, a complete cluster management mechanism is established: automatic container restart, automatic backup, and automatic container scaling; kubernetes designs components for pod, replication controller, and service to manage containers, and provides RESTful interfaces for operations; because this article is mainly about Kubernetes, I will not introduce too much about the theory;
This article deploys Kubernetes to Docker and uses three nodes 192.168.2.143 to both Master and minion nodes. In addition, there are two minion nodes, 192.168.2.144 and 192.168.2.145;

1. The master node starts the etcd container.
Start the etcd container

-- Bind docker run-d -- net = host -- restart = always -- name = etcd-v/var/etcd/data:/var/etcd/data kubernetes/etcd: 2.0.5/usr/local/bin/etcd -- addr = 192.168.2.143: 4001 -- bind-addr = 0.0.0.0: 4001 -- data-dir =/var/etcd/data

Insert the flannel configuration information in etcd and specify that the flannel uses the 10.0.0.0/8 interval.

 docker exec -it etcd etcdctl set /solinx.co/network/config '{"Network":"10.0.0.0/8"}'  

2. Configure flanneld on master, minion1, and minion2 nodes.

Start flanneld, wget-c https://github.com/coreos/flannel/releases/download/v0.5.5/flannel-0.5.5-linux-amd64.tar.gz. /flanneld -- etcd-endpoints = http: // 192.168.2.143: 4001 -- etcd-prefix =/solinx. co/network -- iface = ens33> flannel. log 2> & 1 &

After flannel is started, the IP address set that can be allocated is obtained and stored in/run/flannel/subnet. env. You need to configure available and allocated IP addresses for docker.

Modify the Docker configuration file in Ubuntu

In/lib/systemd/system/docker. added EnvironmentFile =-/etc/default/docker to change ExecStart =/usr/bin/docker-d-H fd: //: execStart =/usr/bin/docker-d-H fd: // $ DOCKER_OPTS source/run/flannel/subnet. env sh-c "echo DOCKER_OPTS =\\" -- bip = $ FLANNEL_SUBNET -- mtu = $ FLANNEL_MTU \\\ ">/etc/default/docker" service docker restart

3. Kubernetes deployment
Download kubernetes.tar.gz to master and minion nodes.

 wget -c https://github.com/kubernetes/kubernetes/releases/download/v1.2.5/kubernetes.tar.gz

Decompress the kubernetes.tar.gz file and find kubernetes-server-linux-amd64.tar.gz in the kubernetes/serverdirectory to decompress the file. In the kubernetes/server/bin directory, find: release;
Import the kubernetes Docker image file to the master node

 docker load -i kube-apiserver.tar docker load -i kube-controller-manager.tar docker load -i kube-scheduler.tar

Here, flannel and kubernetes use the same etcd

Start apiServer on the master node

 docker run -d --name=apiserver --net=host gcr.io/google_containers/kube-apiserver:7bf05b2d35172296e4fbd2604362456f kube-apiserver --insecure-bind-address=192.168.2.143 --service-cluster-ip-range=10.0.0.0/16 --etcd-servers=http://192.168.2.143:4001  

Master node starts ControllerManager

 docker run -d --name=ControllerM gcr.io/google_containers/kube-controller-manager:6c95ef0b57ac9deda34ae1a4a40baa0a kube-controller-manager --master=192.168.2.143:8080

Start Scheduler on the master node.

 docker run -d --name=scheduler gcr.io/google_containers/kube-scheduler:e5342c3d8ced06850af97347daf6ae4b kube-scheduler --master=192.168.2.143:8080

Server startup completed

. /Kubectl-s 192.168.2.143: 8080 version view kubernetes Version information Client version: version. info {Major: "1", Minor: "2", GitVersion: "v1.2.5", GitCommit: "25eb53b54e08877d3789455964b3e97bdd3f3bce", GitTreeState: "clean"} Server Version: version. info {Major: "1", Minor: "2", GitVersion: "v1.2.5", GitCommit: "25eb53b54e08877d3789455964b3e97bdd3f3bce", GitTreeState: "clean "}

View the startup status of the Service Container on the Master node:

 docker ps  

Start kubelet on 143, 144, and 145 minion nodes

Obtain the minion node IP address.

 NODE_IP=`ifconfig ens33 | grep 'inet addr:' | cut -d: -f2 | cut -d' ' -f1` ./kubelet --api-servers=192.168.2.143:8080 --node-ip=$NODE_IP --hostname_override=192.168.2.144 > kubelet.log 2>&1 &

Note: If the two host names are the same, you must use the hostname_override parameter. Otherwise, you must change the host names to different ones;

View node information on the master node

 ./kubectl -s 192.168.2.143:8080 get no

Start kube-proxy on 143, 144, and 145 nodes

 ./kube-proxy --master=192.168.2.143:8080 > proxy.log 2>&1 &

Run the following command to view the pod, replication controller, service, and endpoint:

 ./kubectl -s 192.168.1.143:8080 get po ./kubectl -s 192.168.1.143:8080 get rc ./kubectl -s 192.168.1.143:8080 get svc ./kubectl -s 192.168.1.143:8080 get ep  

References:
Http://kubernetes.io/docs/user-guide/

Kubernetes cluster deployment

OpenStack, Kubernetes, and Mesos

Problems encountered during Kubernetes cluster construction and Solutions

For details about Kubernetes, click here
Kubernetes: click here

This article permanently updates the link address: