**
Swarm架构
**
Swarm作为一个管理Docker集群的工具，首先需要将其部署起来，可以单独将Swarm部署于一个节点。另外，自然需要一个Docker集群，集群上每一个节点均安装有Docker。
Swarm架构中最主要的处理部分自然是Swarm节点，Swarm管理的对象自然是Docker Cluster，Docker Cluster由多个Docker Node组成，而负责给Swarm发送请求的是Docker Client。
Swarm deamon只是一个调度器（Scheduler）加路由器(router)，Swarm自己不运行容器，它只是接受docker客户端发送过来的请求，调度适合的节点来运行容器，这意味着，即使Swarm由于某些原因挂掉了，集群中的节点也会照常运行，当Swarm重新恢复运行之后，它会收集重建集群信息。
**
环境搭建
**
三台主机
server0 172.25.254.56
server1 172.25.254.50
server2 172.25.254.1
安装docker
vim /etc/yum.repos.d/docker.repo
[dockerrepo]
name=Docker Repository
baseurl=https://yum.dockerproject.org/repo/experimental/centos/7/
enabled=1
gpgcheck=0
gpgkey=https://yum.dockerproject.org/gpg
~
yum -y install docker-engine
[root@foundation156 ~]# systemctl start docker
[root@foundation156 ~]# docker images
REPOSITORY TAG IMAGE ID CREATED SIZE
[root@foundation156 ~]# docker --version
Docker version 1.12.6, build 78d1802, experimental
三台主机安装即可
ps （根据前一部分的博客选择rpm包安装也可以）
swarm集群搭建
node0作为初始节点供其他节点加入
Connection to server1 closed.
[root@foundation156 Desktop]# docker swarm init --advertise-addr 172.25.254.56
Swarm initialized: current node (59rkk6ozh5h1ddptl88f52h0v) is now a manager.
To add a worker to this swarm, run the following command:
docker swarm join \
--token SWMTKN-1-589zuow0efnzllasdj1nei7w34p02z6mkm443o3i5avuvn01tj-3ywyld3erpdnf7n4jhbgbzocs \
172.25.254.56:2377
To add a manager to this swarm, run 'docker swarm join-token manager' and follow the instructions.
执行命令后提示其他节点使用系统提示命令加入节点即可
节点加入
[root@server2 docker]# docker swarm join --token SWMTKN-1-589zuow0efnzllasdj1nei7w34p02z6mkm443o3i5avuvn01tj-3ywyld3erpdnf7n4jhbgbzocs 172.25.254.56:2377
This node joined a swarm as a worker.
在node查看端口使用情况
看到使用了2377端口 所以在需要打开防火墙时 需要打开对应的端口
[root@foundation156 Desktop]# netstat -anple |grep docker
tcp 0 0 172.25.254.56:56024 172.25.254.56:2377 ESTABLISHED 0 73800 3548/dockerd
tcp 0 0 127.0.0.1:43323 127.0.0.1:2377 ESTABLISHED 0 72244 3548/dockerd
tcp6 0 0 :::2377 :::* LISTEN 0 70307 3548/dockerd
tcp6 0 0 :::7946 :::* LISTEN 0 70312 3548/dockerd
tcp6 0 0 127.0.0.1:2377 127.0.0.1:43323 ESTABLISHED 0 70309 3548/dockerd
tcp6 0 0 172.25.254.56:2377 172.25.254.56:56024 ESTABLISHED 0 72273 3548/dockerd
tcp6 0 0 172.25.254.56:2377 172.25.254.50:45988 ESTABLISHED 0 73673 3548/dockerd
tcp6 0 0 172.25.254.56:2377 172.25.254.1:51654 ESTABLISHED 0 73669 3548/dockerd
udp6 0 0 :::7946 :::* 0 70313 3548/dockerd
在node0查看是否成为sarm集群
docker info
........
Swarm: active
NodeID: 59rkk6ozh5h1ddptl88f52h0v
Is Manager: true
ClusterID: 8bq7ol10vkzyra3xhlfvuv0t3
Managers: 1
Nodes: 3
Orchestration:
Task History Retention Limit: 5
........
列出所有节点
[root@foundation156 Desktop]# docker node ls
ID HOSTNAME STATUS AVAILABILITY MANAGER STATUS
59rkk6ozh5h1ddptl88f52h0v * foundation156.ilt.example.com Ready Active Leader
6246ae4zn9asricvh2d3uc1qp server1 Ready Active
daugy9sjlh5zet6vnkh0fo1h3 server2 Ready Active
在忘记添加节点命令时 通过以下命令查看
[root@foundation156 Desktop]# docker swarm join-token manager
To add a manager to this swarm, run the following command:
docker swarm join \
--token SWMTKN-1-589zuow0efnzllasdj1nei7w34p02z6mkm443o3i5avuvn01tj-5taetju3kdmcvcunu3h7csm3l \
172.25.254.56:2377
[root@foundation156 Desktop]# docker swarm join-token worker
To add a worker to this swarm, run the following command:
docker swarm join \
--token SWMTKN-1-589zuow0efnzllasdj1nei7w34p02z6mkm443o3i5avuvn01tj-3ywyld3erpdnf7n4jhbgbzocs \
172.25.254.56:2377
目前只有一个主导的节点server0 为了集群高可用并且避免单点故障
提升server1为manager节点 使其也可以管理集群
[root@foundation156 Desktop]# docker node promote server1
Node server1 promoted to a manager in the swarm.
[root@foundation156 Desktop]# docker node ls
ID HOSTNAME STATUS AVAILABILITY MANAGER STATUS
59rkk6ozh5h1ddptl88f52h0v * foundation156.ilt.example.com Ready Active Leader
6246ae4zn9asricvh2d3uc1qp server1 Ready Active Reachable
daugy9sjlh5zet6vnkh0fo1h3 server2 Ready Active
常用命令
docker swarm：集群管理，子命令主要有下面几个。
docker swarm init命令用于初始化一个集群
docker swarm join命令用于加入一个现有集群
docker swarm leave命令由于离开集群
有些时候需要维护一个节点，此时此节点可能会网络断开或者需要关机，造成节点上服务不可用。使用 docker node update –availability drain 将节点下线，swarm 会将当前节点上的容器关闭并在其他节点上启动。当维护完成，需要上线是，将节点状态修改为 active状态即可,命令如下：docker node update –availability active
运行服务测试集群性能
在物理机pull nginx镜像 本纪可以使用 但是虚拟机上网下载太满 所以将物理机镜像保存为tar文件scp让虚拟机使用即可
[root@foundation156 Desktop]# docker save -o nginx.tar nginx
[root@foundation156 Desktop]# ll nginx.tar
-rw------- 1 root root 112703488 Aug 20 20:18 nginx.tar
[root@foundation156 Desktop]# scp nginx.tar root@server1:/mnt/
nginx.tar 100% 107MB 107.5MB/s 00:01
[root@foundation156 Desktop]# scp nginx.tar root@server2:/mnt/
nginx.tar 100% 107MB 107.5MB/s 00:01
[root@foundation156 Desktop]# docker service rm nginx
nginx
虚拟机导入镜像即可
[root@server2 docker]# vim /etc/sysconfig/network-scripts/ifcfg-ens3
[root@server2 docker]# cd /mnt/
[root@server2 mnt]# docker load -i nginx.tar
cdb3f9544e4c: Loading layer 58.44 MB/58.44 MB
a8c4aeeaa045: Loading layer 54.24 MB/54.24 MB
08d25fa0442e: Loading layer 3.584 kB/3.584 kB
Loaded image: nginx:latest
[root@server2 mnt]# docker images
REPOSITORY TAG IMAGE ID CREATED SIZE
nginx latest c82521676580 3 weeks ago 109 MB
[root@foundation156 Desktop]# docker pull nginx
Using default tag: latest
latest: Pulling from library/nginx
be8881be8156: Pull complete
32d9726baeef: Pull complete
87e5e6f71297: Pull complete
Digest: sha256:d85914d547a6c92faa39ce7058bd7529baacab7e0cd4255442b04577c4d1f424
Status: Downloaded newer image for nginx:latest
[root@foundation156 Desktop]# docker images nginx
REPOSITORY TAG IMAGE ID CREATED SIZE
nginx latest c82521676580 3 weeks ago 109 MB
[root@foundation156 Desktop]# docker service create --replicas 3 -p 80:80 --name nginx nginx:latest
244icxsoy3n8g5gh4otdio0pp
[root@foundation156 Desktop]# docker service ls
ID NAME REPLICAS IMAGE COMMAND
244icxsoy3n8 nginx 1/3 nginx:latest
查看物理机nginx服务详细信息
[root@foundation156 Desktop]# docker service ps nginx
ID NAME IMAGE NODE DESIRED STATE CURRENT STATE ERROR
cqawk4av2vj2579lsc7n8udse nginx.1 nginx:latest server0 Running Running 3 seconds ago
1rvhljwyvajyklqkbjgn7fbt8 nginx.2 nginx:latest server2 Running Preparing 8 seconds ago
bhkx7if20gcwd8wrgkq7aqcqr nginx.3 nginx:latest server1 Running Running 6 seconds ago
查看节点的进程
[root@server1 mnt]# docker ps
CONTAINER ID IMAGE COMMAND CREATED STATUS PORTS NAMES
7aade034cd8b nginx:latest "nginx -g 'daemon ..." 2 minutes ago Up 2 minutes 80/tcp nginx.2.bn25l3rukkufe34ry7kkzu06h
浏览器测试即可
使用 visualzer进行测试
安装镜像
[root@server1 ~]# docker run -it -d -p 8080:8080 -v /var/run/docker.sock:/var/run/docker.sock dockersamples/visualizer
e6bc05f62386ea0c5730dbb6c5393224e81f61e831a2931c14cbff04d1c79011
浏览器测试即可
关闭server后
服务漂移至server3
[root@server3 ~]# docker ps
CONTAINER ID IMAGE COMMAND CREATED STATUS PORTS NAMES
6e599b133fd8 nginx:latest "nginx -g 'daemon ..." 5 minutes ago Up 4 minutes 80/tcp nginx.1.8yy6rf4g565xqeb6q50y8hroa
e7ef3ec12304 nginx:latest "nginx -g 'daemon ..." 33 minutes ago Up 33 minutes 80/tcp nginx.2.qq3si9jorvrml2dq5ftkvshny
再测试负载均衡
成功
[root@foundation15 docker]# for i in {1..10};do curl 172.25.15.20; done
server1
server3
server2
server1
server3
server2
server1
server3
server2
server1批量化生成nginx
[root@server1 ~]# docker service scale nginx=30
批量更新 这一步再生产中作用较大
测试再nginx主机上更新2048镜像
[root@server1 ~]# docker service update --image westos.org/game2048 --update-parallelism 3 --update-delay 10s nginx
nginx等待更新即可
看到30台已经更新完成
[root@server1 ~]# docker service ls
ID NAME MODE REPLICAS IMAGE
1mj2yiaoo9ok viz replicated 1/1 dockersamples/visualizer:latest
p16i2s6m74zg nginx replicated 30/30 westos.org/game2048:latest