【实战】Docker Machine + Compose + Swarm
【编者的话】本文的案例结合了Docker的三大编排工具Docker Machine、Compose与Swarm，值得读者们参阅。Docker现在已经推出了一些新的命令行工具，这些工具可以简单的编排（orchestration ）Docker实例、集群以及容器管理。它们是：
[*] Docker Machine - 让你轻松部署Docker实例到很多不同的平台。
[*] Docker Compose - Fig工具的替代品。
[*] Docker Swarm - Docker众实例的原生集群。
这三种技术中，Swarm目前不适合在生产中使用，因此在这篇文章中我不会讲关于它的太多细节。
Docker Machine
对于直接下载预编译的二进制文件来说，我决定使用Homebrew（OS X的管理包工具）：
# Make sure everything is up-to-date
brew update
brew doctor
brew cask update
brew cask doctor
# install docker-machine
brew cask install docker-machine
这样就安装了docker-machine。
$docker-machine -v
docker-machine version 0.1.0
$docker-machine ls
NAME ACTIVE DRIVER STATE URL SWARM
$
我已经安装了VirtualBox，并且要创建一个叫“testing”的虚拟机：
$docker-machine create --driver virtualbox testing
INFO Creating SSH key...
INFO Creating VirtualBox VM...
INFO Starting VirtualBox VM...
INFO Waiting for VM to start...
INFO "testing" has been created and is now the active machine.
INFO To point your Docker client at it, run this in your shell: $(docker-machine env testing)
docker-machine使用几个命令来帮助你连接到本地安装的Docker客户端：
$docker-machine env testing
export DOCKER_TLS_VERIFY=yes
export DOCKER_CERT_PATH=/Users/russ/.docker/machine/machines/testing
export DOCKER_HOST=tcp://192.168.99.100:2376
$docker-machine config testing
--tls --tlscacert=/Users/russ/.docker/machine/machines/testing/ca.pem --tlscert=/Users/russ/.docker/machine/machines/testing/cert.pem --tlskey=/Users/russ/.docker/machine/machines/testing/key.pem -H="tcp://192.168.99.100:2376
我现在启用了一个虚拟机并准备使用Docker。
$docker-machine ls
NAME ACTIVE DRIVER STATE URL SWARM
testing * virtualbox Running tcp://192.168.99.100:2376
和其他新安装的一样，让我们运行一个“Hello World”：
$docker $(docker-machine config testing) run busybox echo hello world
Unable to find image 'busybox:latest' locally
511136ea3c5a: Pull complete
df7546f9f060: Pull complete
ea13149945cb: Pull complete
4986bf8c1536: Pull complete
busybox:latest: The image you are pulling has been verified. Important: image verification is a tech preview feature and should not be relied on to provide security.
Status: Downloaded newer image for busybox:latest
hello world
最后，你可以使用docker-machie ssh machine-name命令SSH到虚拟机：
$docker-machine ssh testing
Boot2Docker version 1.5.0, build master : a66bce5 - Tue Feb 10 23:31:27 UTC 2015
Docker version 1.5.0, build a8a31ef
docker@testing:~$ uname -a
Linux testing 3.18.5-tinycore64 #1 SMP Sun Feb 1 06:02:30 UTC 2015 x86_64 GNU/Linux
docker@testing:~$ cat /etc/*release
NAME=Boot2Docker
VERSION=1.5.0
ID=boot2docker
ID_LIKE=tcl
VERSION_ID=1.5.0
PRETTY_NAME="Boot2Docker 1.5.0 (TCL 5.4); master : a66bce5 - Tue Feb 10 23:31:27 UTC 2015"
ANSI_COLOR="1;34"
HOME_URL="http://boot2docker.io"
SUPPORT_URL="https://github.com/boot2docker/boot2docker"
BUG_REPORT_URL="https://github.com/boot2docker/boot2docker/issues"
docker@testing:$ exit
$
太棒了，我现在有一个虚拟机运行在我的电脑上，接下来呢？
设计docker-machine就是和以下公有和私有的云服务提供商（以后会添加更多）一起使用的。
[*] Amazon EC2
[*] Microsoft Azure
[*] Digital Ocean
[*] Google Compute Engine
[*] Rackspace
[*] SoftLayer
[*] OpenStack
[*] VMWare vCloud Air
[*] VMWare vSphere
让我们使用docker-machine来启用一个Digital Ocean的实例。你需要生成一个Personal Access Token。一旦用token启用机子就会像下面所示一样：
$docker-machine create \
→ --driver digitalocean \
→ --digitalocean-access-token cdb81ed0575b5a8d37cea0d06c9690daa074b1276892fc8473bdda97eb7c65ae \
→ dotesting
INFO Creating SSH key...
INFO Creating Digital Ocean droplet...
INFO Waiting for SSH...
INFO Configuring Machine...
INFO "dotesting" has been created and is now the active machine.
INFO To point your Docker client at it, run this in your shell: $(docker-machine env dotesting)
那么发生了什么呢？docker-machine访问我的Digital Ocean账户通过API并且启用了以下配置的实例：
[*] OS = Ubuntu 14.04 x64
[*] RAM = 512MB
[*] HDD = 20GB SSD
[*] Region = NYC3
这些默认的配置可以通过提供更多的选项被修改，运行docker-machine create --help获取帮助查看所有带例子的选项。
一旦实例开启，docker-machine通过SSH连接到安装、配置以及开启的最新的Docker上。
所以，我们现在有两台Machines，一个在本地，一个在Digital Ocean上。
$docker-machine ls
NAME ACTIVE DRIVER STATE URL SWARM
dotesting * digitalocean Running tcp://45.55.134.248:2376
testing virtualbox Running tcp://192.168.99.100:2376
让我们再次运行“Hello World”，但是这次使用刚才启动的那个实例：
$docker $(docker-machine config dotesting) run busybox echo hello world
Unable to find image 'busybox:latest' locally
511136ea3c5a: Pull complete
df7546f9f060: Pull complete
ea13149945cb: Pull complete
4986bf8c1536: Pull complete
busybox:latest: The image you are pulling has been verified. Important: image verification is a tech preview feature and should not be relied on to provide security.
Status: Downloaded newer image for busybox:latest
hello world
并且SSH到那个机器中：
$docker-machine ssh dotesting
Welcome to Ubuntu 14.04.1 LTS (GNU/Linux 3.13.0-43-generic x86_64)
Documentation:https://help.ubuntu.com/
System information as of Sat Mar 21 07:24:02 EDT 2015
System load:0.43 Processes: 72
Usage of /: 11.4% of 19.56GB Users logged in: 0
Memory usage: 12% IP address for eth0: 45.55.134.248
Swap usage: 0% IP address for docker0: 172.17.42.1
Graph this data and manage this system at:
https://landscape.canonical.com/
root@dotesting:~# docker images
REPOSITORY TAG IMAGE ID CREATED VIRTUAL SIZE
busybox latest 4986bf8c1536 11 weeks ago 2.433 MB
root@dotesting:~# docker ps -a
CONTAINER ID IMAGE COMMAND CREATED STATUS PORTS NAMES
b8a83077d858 busybox:latest "echo hello world" 4 minutes ago Exited (0) 4 minutes ago kickass_almeida
root@dotesting:~# exit
logout
$
最终，你可以使用docker-mashie stop machine-name和docker-mashie rm machine-name来停止和移除machines。请注意当使用rm时，是不会提示你是否确定删除。
$docker-machine ls
NAME ACTIVE DRIVER STATE URL SWARM
dotesting * digitalocean Running tcp://45.55.134.248:2376
testing virtualbox Running tcp://192.168.99.100:2376
$docker-machine stop dotesting
$docker-machine ls
NAME ACTIVE DRIVER STATE URL SWARM
dotesting * digitalocean Stopped tcp://45.55.134.248:2376
testing virtualbox Running tcp://192.168.99.100:2376
$docker-machine rm dotesting
$docker-machine ls
NAME ACTIVE DRIVER STATE URL SWARM
testing virtualbox Running tcp://192.168.99.100:2376
$
总结，以上就是docker-machie的总览。正如你看到的，它确实很方便在很多不同的供应商中来引导Docker服务实例，并且使用一个本地machine命令就可以操控他们。
Docker Compose
Docker开始充满生机是因为有了Fig，这是我曾在以前的文章中写到过，当前版本并没有添加太多的新功能，但它开始奠定了与docker-swarm工作的基础，单击这里查看详细日志。
像docker-machine一样，我使用Homebrew公式安装docker-compose。
$brew install docker-compose
==> Downloading https://homebrew.bintray.com/bottles/fig-1.1.0.yosemite.bottle.1.tar.gz############################################################ 100.0%==> Pouring fig-1.1.0.yosemite.bottle.1.tar.gz
==> Caveats
Bash completion has been installed to:
/usr/local/etc/bash_completion.d
==> Summary
/usr/local/Cellar/fig/1.1.0: 186 files, 2.2M
$
然后，使用docker-machine创建一个Docker服务实例：
$docker-machine create --driver virtualbox compose
INFO Creating SSH key...
INFO Creating VirtualBox VM...
INFO Starting VirtualBox VM...
INFO Waiting for VM to start...
INFO "compose" has been created and is now the active machine.
INFO To point your Docker client at it, run this in your shell: $(docker-machine env compose)
因为docker-compose不直接与docker-machine交互，我们需要告诉docker客户端那些刚刚启动的服务器实例的详细信息。
$$(docker-machine env compose)
此命令注入所需Docker客户端的环境变量来连接到服务实例，要看到他们，你只需运行docker-machine env machine-name。
$docker-machine env compose
export DOCKER_TLS_VERIFY=yes
export DOCKER_CERT_PATH=/Users/russ/.docker/machine/machines/compose
export DOCKER_HOST=tcp://192.168.99.100:2376
往后它就像Fig一样，除了fig.yml文件现在应该改为docker-compose.yml，在我以前的博文里有一个fig.yml文件描述：
web:
image: russmckendrick/nginx-php
volumes:
- ./web:/var/www/html/
ports:
- 80:80
environment:
PHP_POOL: mywebsite
links:
- db:db
db:
image: russmckendrick/mariadb
ports:
- 3306
privileged: true
environment:
MYSQL_ROOT_PASSWORD: wibble
MYSQL_DATABASE: wibble
MYSQL_USER: wibble
MYSQL_PASSWORD: wibble
它启用两个容器并且把它们连接到一起，还有在Nginx容器内的/var/www/html被挂载到host的./web文件夹下。我准备运行docker-compose命令的文件夹的结构是这样的：
$tree -a
.
├── \docker-compose.yml
└── \web
└── \index.php
1 directory, 2 files
开始我要拉取需要启用的镜像，你可以忽略此部分。
$docker-compose pull
Pulling db (russmckendrick/mariadb:latest)...
Pulling web (russmckendrick/nginx-php:latest)...
现在镜像已经被拉取下来，是时候开启容器了：
$docker-compose up -d
Creating example_db_1...
Creating example_web_1...
我们现在有了两个正在运行的容器：
$docker-compose ps
Name Command State Ports
----------------------------------------------------------------
example_db_1 /usr/local/bin/run Up 0.0.0.0:49154->3306/tcp
example_web_1 /usr/local/bin/run Up 0.0.0.0:80->80/tcp
你也可以打开浏览器：
open http://$(docker-machine ip)
在我的例子中我看到PHPinfo()页面，如下：
一旦容器开启，你可以使用docker exec来连接到容器内部：
$docker exec -it example_web_1 bash
# ps aux
USER PID %CPU %MEM VSZ RSS TTY STAT START TIME COMMAND
root 10.21.5 115200 15360 ? Ss 13:59 0:01 /usr/bin/python /usr/bin/supervisord -n
root 160.03.2 382876 33624 ? S 13:59 0:00 php-fpm: master process (/etc/php-fpm.conf)
root 170.00.2 1100162096 ? Ss 13:59 0:00 nginx: master process nginx
nginx 180.00.5 1104725568 ? S 13:59 0:00 nginx: worker process
webserv+ 190.01.5 383132 16284 ? S 13:59 0:00 php-fpm: pool mywebsite
webserv+ 200.00.8 3828768848 ? S 13:59 0:00 php-fpm: pool mywebsite
webserv+ 210.00.8 3828768848 ? S 13:59 0:00 php-fpm: pool mywebsite
webserv+ 220.00.8 3828768848 ? S 13:59 0:00 php-fpm: pool mywebsite
webserv+ 230.00.8 3828768852 ? S 13:59 0:00 php-fpm: pool mywebsite
root 950.00.4915404740 ? Ss 13:59 0:00 /usr/libexec/postfix/master -w
postfix 970.00.6917126508 ? S 13:59 0:00 qmgr -l -t unix -u
postfix 2000.00.6916446232 ? S 14:05 0:00 pickup -l -t unix -u
root 2342.30.2117482968 ? S 14:07 0:00 bash
root 2501.01.1 110012 11616 ? S 14:07 0:00 nginx
root 2510.00.2197562212 ? R+ 14:07 0:00 ps aux
# exit
exit
$
最后你可以停止以及移除容器，当然还有Docker实例：
$docker-compose stop && docker-compose rm --force
Stopping example_web_1...
Stopping example_db_1...
Going to remove example_web_1, example_db_1
Removing example_db_1...
Removing example_web_1...
$docker-machine rm compose
$docker-machine ls
NAME ACTIVE DRIVER STATE URL SWARM
$
Docker Swarm
在进一步讨论之前，先看下官方文档的警告:
警告：Swarm当前是beta版本，因此后期可能会有变化。我们还不推荐在生产环境中使用。 现在让我们使用Homebrew来安装docker-swarm：
$brew install docker-swarm
==> Downloading https://homebrew.bintray.com/bottles/docker-swarm-0.1.0.yosemite.bottle.tar.gz############################################################ 100.0%==> Pouring docker-swarm-0.1.0.yosemite.bottle.tar.gz
/usr/local/Cellar/docker-swarm/0.1.0: 4 files, 8.7M
因为我们已经安装了docker-machine，我将要使用它在本地创建一个集群。首先，我们需要启动一个实例并运行swarm容器：
$docker-machine ls
NAME ACTIVE DRIVER STATE URL SWARM
$docker-machine create -d virtualbox local
INFO Creating SSH key...
INFO Creating VirtualBox VM...
INFO Starting VirtualBox VM...
INFO Waiting for VM to start...
INFO "local" has been created and is now the active machine.
INFO To point your Docker client at it, run this in your shell: $(docker-machine env local)
$ $(docker-machine env local)
$docker run swarm create
Unable to find image 'swarm:latest' locally
511136ea3c5a: Pull complete
ae115241d78a: Pull complete
f49087514537: Pull complete
fff73787bd9f: Pull complete
97c8f6e912d7: Pull complete
33f9d1e808cf: Pull complete
62860d7acc87: Pull complete
bf8b6923851d: Pull complete
swarm:latest: The image you are pulling has been verified. Important: image verification is a tech preview feature and should not be relied on to provide security.
Status: Downloaded newer image for swarm:latest
63e7a1adb607ce4db056a29b1f5d30cf
$
正如你所见，当容器启动时，我得到了一个token：63e7a1adb607ce4db056a29b1f5d30cf，我将要用它来添加更多的节点，但是首先我们需要创建一个Swarm master：
$docker-machine create \
→ -d virtualbox \
→ --swarm \
→ --swarm-master \
→ --swarm-discovery token://63e7a1adb607ce4db056a29b1f5d30cf \
→ swarm-master
INFO Creating SSH key...
INFO Creating VirtualBox VM...
INFO Starting VirtualBox VM...
INFO Waiting for VM to start...
INFO Configuring Swarm...
INFO "swarm-master" has been created and is now the active machine.
INFO To point your Docker client at it, run this in your shell: $(docker-machine env swarm-master)
然后，我们需要连接Docker客户端到Swarm上，这就需要将--swarm添加到$(docker-machine env machine-name)命令上：
$$(docker-machine env --swarm swarm-master)
现在让我们添加另一个节点：
$docker-machine create \
→ -d virtualbox \
→ --swarm \
→ --swarm-discovery token://63e7a1adb607ce4db056a29b1f5d30cf\
→ swarm-node-00
INFO Creating SSH key...
INFO Creating VirtualBox VM...
INFO Starting VirtualBox VM...
INFO Waiting for VM to start...
INFO Configuring Swarm...
INFO "swarm-node-00" has been created and is now the active machine.
我们现在有了2个节点的集群 - “swarm-master”：
$docker-machine ls
NAME ACTIVE DRIVER STATE URL SWARM
local virtualbox Running tcp://192.168.99.100:2376
swarm-master virtualbox Running tcp://192.168.99.101:2376 swarm-master (master)
swarm-node-00 * virtualbox Running tcp://192.168.99.102:2376 swarm-master
使用docker info来获取更多有关集群的信息：
$docker info
Containers: 3
Nodes: 2
swarm-master: 192.168.99.101:2376
└ Containers: 2
└ Reserved CPUs: 0 / 4
└ Reserved Memory: 0 B / 999.9 MiB
swarm-node-00: 192.168.99.102:2376
└ Containers: 1
└ Reserved CPUs: 0 / 4
└ Reserved Memory: 0 B / 999.9 MiB
太棒了，但这意味着什么？
让我们拉取一些镜像：
$docker -H 192.168.99.101:2376 pull redis
$docker -H 192.168.99.102:2376 pull mysql
注意到我是如何在“swarm-master”上拉取redis镜像以及在swarm-node-00上拉取mysql的，现在我可以保证容器只在有镜像的那个节点上启用：
$docker run -d --name redis1 -e affinity:image==redis redis
af66148bbbc8dcd799d82448dfd133b968d34eb7066a353108bf909ea3324a58
$docker run -d --name mysql -e affinity:image==mysql -e MYSQL_ROOT_PASSWORD=mysecretpassword -d mysql
70b2d93d6f83aa99f5ad4ebe5037e228a491a4b570609840f3f4be9780c33587
$docker ps
CONTAINER ID IMAGE COMMAND CREATED STATUS PORTS NAMES
70b2d93d6f83 mysql:latest "/entrypoint.sh mysq 3 seconds ago Up Less than a second 3306/tcp swarm-node-00/mysql
af66148bbbc8 redis:latest "/entrypoint.sh redi 2 minutes ago Up 2 minutes 6379/tcp swarm-master/redis1
另一个例子是使用节点的端口，让我们在两个节点上都拉取我的nginx-php镜像：
$docker -H 192.168.99.101:2376 pull russmckendrick/nginx-php
$docker -H 192.168.99.102:2376 pull russmckendrick/nginx-php
现在，让我们启用一个容器并绑定到80端口。
$docker run -d -p 80:80 russmckendrick/nginx-php
2d066b2ccf28d2a1fa9edad8ac7b065266f29ecb49a8753b972780051ff83587
再有：
$docker run -d -p 80:80 russmckendrick/nginx-php
40f5fee257bb2546a639a5dc5c2d30f8fa0ac169145e431c534f85d8db51357f
你会说这没什么特别的啊？正常来说，当试图启动第二个容器时，你会得到如下信息因为你不用将同一个端口绑定到两个容器上：
$docker run -d -p 80:80 russmckendrick/nginx-php
FATA Error response from daemon: unable to find a node with port 80 available
然而，在集群的情况下，因为Docker知道集群节点运行的是什么以及哪些端口是在使用的。Docker可以简单地通过Swarm在“swarm-node-00”上启动容器并且它知道“swarm-master”已经使用了80端口：
$docker ps
CONTAINER ID IMAGE COMMAND CREATED STATUS PORTS NAMES
40f5fee257bb russmckendrick/nginx-php:latest "/usr/local/bin/run" 4 seconds ago Up Less than a second 192.168.99.101:80->80/tcp swarm-master/elated_einstein
2d066b2ccf28 russmckendrick/nginx-php:latest "/usr/local/bin/run" 8 seconds ago Up Less than a second 192.168.99.102:80->80/tcp swarm-node-00/drunk_mestorf
70b2d93d6f83 mysql:latest "/entrypoint.sh mysq 26 minutes ago Up 26 minutes 3306/tcp swarm-node-00/mysql
af66148bbbc8 redis:latest "/entrypoint.sh redi 29 minutes ago Up 29 minutes 6379/tcp swarm-master/redis1
所有这一切都没有提示或特殊的命令行，它的帮助仅仅是用它来做到这些。
正如你所看到的，docker-swarm仍然有非常大的发展潜力，但也有一些不如意的地方，如容器不能够跨节点通讯。然而，伴随着socketplane.io（他们使用Open vSwitch开发了一个基于软件定义网络解决方案的容器）的加入，我想用不了多长时间这个问题就能得到解决。
最后，让我们删除在运行的实例：
$docker-machine rm local swarm-master swarm-node-00
就这样吧，期待在未来的几个月这些工具的更新，我也会进一步跟进。
原文：Docker Machine, Compose & Swarm （翻译：田浩浩 校对：李颖杰）
页:
[1]