VM1: 129.107.126.123
VM2: 129.107.126.142
container c1放在vm1上，container c2放在vm2上。etcd放在vm1上。vm1和vm2加入一个overlay网络。
1, run etcd
安装etcd不谈
$ etcd --advertise-client-urls 'http://129.107.126.123:2379' --listen-client-urls 'http://0.0.0.0:2379'
本地检查：
$ etcdctl member list
ce2a822cea30bfca: name=default peerURLs=http://localhost:2380,http://localhost:7001 clientURLs=http://129.107.126.123:2379
检查功能是否正确,能否正确设置和获取
$ etcdctl mk key value
$ etcdctl get key
value
远程检查：
$ etcdctl --endpoint "http://129.107.126.123:2379" get key
value
$ etcdctl -endpoint http://192.168.15.232:2379 mk key2 value2
value2
$ etcdctl -endpoint http://192.168.15.232:2379 get key2
value2
2, add daemon.json to /etc/docker in VM1
$ sudo vim /etc/docker/daemon.json
{
"cluster-store":"etcd://129.107.126.123:2379",
"cluster-advertise":"129.107.126.123:2376"
}
VM2:
{
"cluster-store":"etcd://129.107.126.123:2379",
"cluster-advertise":"129.107.126.142:2376"
}
其中cluster-store指向etcd所在设备ip, cluster-advertise内填写本机ip地址
$ sudo service docker restart
$ sudo docker info | grep "Cluster"
WARNING: No swap limit support
Cluster Store: etcd://129.107.126.123:2379
Cluster Advertise: 129.107.126.123:2376
说明集群信息正确配置进入了docker deamon
3, 创建overlay network
$ sudo docker network create -d overlay myoverlay-1
17c7751410e7305e88e205521775e3bf99dc2406c80b1219b72274fa884bd3c3
$ sudo docker network ls
NETWORK ID NAME DRIVER SCOPE
da561e6ef90c bridge bridge local
26362372b64b host host local
17c7751410e7 myoverlay-1 overlay global
a72c2bebcc6a none null local
可以看到网络已建立，是global类型
在其他vm上执行，docker network ls，也能看到
4，容器加入现有network
加入网络 myoverlay-1
$ sudo docker run -t -i --name overlay-docker1 --net myoverlay-1 ubuntu:16.10 /bin/bash
已经在运行的容器也可以加入网络,通过执行命令:
$ docker run -id --name t233_t1 busybox sh
$ docker network connect myoverlay-1 t233_out
分别在vm1和vm2上面建立2个container，加入network myoverlay-1
container c1 on vm1: 10.0.0.2
container c2 on vm2: 10.0.0.3
可以ping通
done