1. Main official guidance document
http://kubernetes.io/docs/admin/high-availability/#establishing-a-redundant-reliable-data-storage-layer
2. 192.168.200.12 node
192.168.200.14 master
192.168.200.15 master
192.168.200.16 master
172.16.59.5>>172.16.56.56>> 200.14
3. Install kubelet on the master, the configuration is as follows, pay attention to --register-node=false, you can make the kubelet node not automatically register itself with the apiserver, --config=/etc/kubernetes/manifests The directory monitored by kubelet can be automatically started inside pod's yaml file
###
# kubernetes kubelet (minion) config
# The address for the info server to serve on (set to 0.0.0.0 or "" for all interfaces)
KUBELET_ADDRESS="--address=127.0.0.1"
# The port for the info server to serve on
# KUBELET_PORT="--port=10250"
# You may leave this blank to use the actual hostname
KUBELET_HOSTNAME="--hostname-override=192.168.200.14"
# location of the api-server
KUBELET_API_SERVER=""
# pod infrastructure container
KUBELET_POD_INFRA_CONTAINER="--pod-infra-container-image=index.tenxcloud.com/google_containers/pause:0.1.0"
# Add your own!
KUBELET_ARGS="--config=/etc/kubernetes/manifests --register-node=false"
4. Configure etcd cluster
5. Create pod version of apiserver, controller-manager and scheduler
(1) The mirror image in the official document cannot be pulled down, and the mirror image is downloaded in the mirror square of Speed Cloud, and it cannot run continuously. So I decided to package the image with dockerfile myself
The relevant dockerfile is:
1.apiserver：
FROM index.tenxcloud.com/google_containers/kube-apiserver:v1.2.0
MAINTAINER msxu [email protected]
CMD ["/usr/local/bin/kube-apiserver","--service-cluster-ip-range=10.254.0.0/16","--insecure-bind-address=0.0.0.0","--etcd-servers=http://192.168.200.14:2379,http://192.168.200.1
5:2379,http://192.168.200.16:2379","--admission-control=AlwaysAdmit"]
2.controller-manager:注意参数--leader-elect=true，用来做controller-manager和scheduler的选主的
FROM index.tenxcloud.com/google_containers/kube-controller-manager:v1.2.2
MAINTAINER msxu [email protected]
CMD ["/usr/local/bin/kube-controller-manager","--master=192.168.200.14:8081","--cluster-cidr=10.245.0.0/16","--leader-elect=true"]
3.scheduler:
FROM index.tenxcloud.com/google_containers/kube-scheduler:v1.2.2
MAINTAINER msxu [email protected]
CMD ["/usr/local/bin/kube-scheduler","--master=192.168.200.14:8081","--leader-elect=true"]
（2）相关的pod文件，放在/etc/kubernetes/manifest目录下，kubelet会自动启动相关的pod
1.kube-apiserver.yaml文件
apiVersion: v1
kind: Pod
metadata:
name: kube-apiserver
spec:
hostNetwork: true
containers:
- name: kube-apiserver
image: index.tenxcloud.com/google_containers/kube-apiserver:msxu0.3.5
ports:
- containerPort: 8080
hostPort: 8080
name: local
volumeMounts:
- mountPath: /var/log/kube-apiserver.log
name: logfile
volumes:
- hostPath:
path: /var/log/kube-apiserver.log
name: logfile
2.kube-scheduler.yaml文件
apiVersion: v1
kind: Pod
metadata:
name: kube-scheduler
spec:
hostNetwork: true
containers:
- name: kube-scheduler
image: index.tenxcloud.com/google_containers/kube-scheduler:msxu0.3.2
livenessProbe:
httpGet:
path: /healthz
port: 10251
initialDelaySeconds: 15
timeoutSeconds: 1
volumeMounts:
- mountPath: /var/log/kube-scheduler.log
name: logfile
volumes:
- hostPath:
path: /var/log/kube-scheduler.log
name: logfile
3.kube-controller-manager.yam文件
apiVersion: v1
kind: Pod
metadata:
name: kube-controller-manager
spec:
hostNetwork: true
containers:
- name: kube-controller-manager
image: index.tenxcloud.com/google_containers/kube-controller-manager:msxu0.3.2
livenessProbe:
httpGet:
path: /healthz
port: 10252
initialDelaySeconds: 15
timeoutSeconds: 1
volumeMounts:
- mountPath: /var/log/kube-controller-manager.log
name: logfile
volumes:
- hostPath:
path: /var/log/kube-controller-manager.log
name: logfile
6.配置nginx服务器
安装略，配置如下：/usr/local/nginx-1.5.1/conf/nginx.conf
7.测试：主要验证当其中一台master机器挂了，其他master能否接管，执行pod保持能力、弹性调度等。。。测试的pod的yaml文件如下：
apiVersion: v1
kind: ReplicationController
metadata:
name: redis-slave
labels:
name: redis-slave
spec:
replicas: 1
selector:
name: redis-slave
template:
metadata:
labels:
name: redis-slave
spec:
containers:
- name: slave
image: docker.io/kubeguide/guestbook-redis-slave
ports:
- containerPort: 6379
http://blog.csdn.net/u012214983/article/details/52267476