【實戰】Docker Machine + Compose + Swarm
本文的案例結合了Docker的三大編排工具Docker Machine、Compose與Swarm,值得讀者們參閱。
Docker現在已經推出了一些新的命令行工具,這些工具可以簡單的編排(orchestration )Docker實例、集群以及容器管理。它們是:
- Docker Machine - 讓你輕松部署Docker實例到很多不同的平臺。
- Docker Compose - Fig工具的替代品。
- Docker Swarm - Docker眾實例的原生集群。 </ul>
這三種技術中,Swarm目前不適合在生產中使用,因此在這篇文章中我不會講關于它的太多細節。
Docker Machine
對于直接下載預編譯的二進制文件來說,我決定使用 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 doctorinstall docker-machine
brew cask install docker-machine</pre>
這樣就安裝了docker-machine。$ docker-machine -v docker-machine version 0.1.0 $ docker-machine ls NAME ACTIVE DRIVER STATE URL SWARM $
我已經安裝了 VirtualBox,并且要創建一個叫“testing”的虛擬機:
$ docker-machine create --driver virtualbox testing INFO[0000] Creating SSH key... INFO[0000] Creating VirtualBox VM... INFO[0006] Starting VirtualBox VM... INFO[0006] Waiting for VM to start... INFO[0038] "testing" has been created and is now the active machine. INFO[0038] To point your Docker client at it, run this in your shell: $(docker-machine env testing)
docker-machine使用幾個命令來幫助你連接到本地安裝的Docker客戶端:
$ 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
我現在啟用了一個虛擬機并準備使用Docker。
$ docker-machine ls NAME ACTIVE DRIVER STATE URL SWARM testing * virtualbox Running tcp://192.168.99.100:2376
和其他新安裝的一樣,讓我們運行一個“Hello World”:
$ docker $(docker-machine config testing) run busybox echo hello world Unable to find image 'busybox:latest' locally 511136ea3c5a: Pull complete df7546f9f060: Pull complete ea13149945cb: Pull complete 4986bf8c1536: Pull complete busybox:latest: The image you are pulling has been verified. Important: image verification is a tech preview feature and should not be relied on to provide security.Status: Downloaded newer image for busybox:latest hello world</pre>
最后,你可以使用docker-machie ssh machine-name命令SSH到虛擬機:$ 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,我現在有一個虛擬機運行在我的電腦上,接下來呢?
設計docker-machine就是和以下公有和私有的云服務提供商(以后會添加更多)一起使用的。
- Amazon EC2
- Microsoft Azure
- Digital Ocean
- Google Compute Engine
- Rackspace
- SoftLayer
- OpenStack
- VMWare vCloud Air
- VMWare vSphere </ul>
讓我們使用docker-machine來啟用一個Digital Ocean的實例。你需要生成一個Personal Access Token。一旦用token啟用機子就會像下面所示一樣:
$ docker-machine create \ → --driver digitalocean \ → --digitalocean-access-token cdb81ed0575b5a8d37cea0d06c9690daa074b1276892fc8473bdda97eb7c65ae \ → dotesting INFO[0000] Creating SSH key... INFO[0000] Creating Digital Ocean droplet... INFO[0004] Waiting for SSH... INFO[0071] Configuring Machine... INFO[0120] "dotesting" has been created and is now the active machine. INFO[0120] To point your Docker client at it, run this in your shell: $(docker-machine env dotesting)
那么發生了什么呢?docker-machine訪問我的Digital Ocean賬戶通過API并且啟用了以下配置的實例:
- OS = Ubuntu 14.04 x64
- RAM = 512MB
- HDD = 20GB SSD
- Region = NYC3 </ul>
這些默認的配置可以通過提供更多的選項被修改,運行docker-machine create --help獲取幫助查看所有帶例子的選項。
一旦實例開啟,docker-machine通過SSH連接到安裝、配置以及開啟的最新的Docker上。
所以,我們現在有兩臺Machines,一個在本地,一個在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
讓我們再次運行“Hello World”,但是這次使用剛才啟動的那個實例:
$ docker $(docker-machine config dotesting) run busybox echo hello world Unable to find image 'busybox:latest' locally 511136ea3c5a: Pull complete df7546f9f060: Pull complete ea13149945cb: Pull complete 4986bf8c1536: Pull complete busybox:latest: The image you are pulling has been verified. Important: image verification is a tech preview feature and should not be relied on to provide security.Status: Downloaded newer image for busybox:latest hello world</pre>
并且SSH到那個機器中:$ 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 $ </pre>
最終,你可以使用docker-mashie stop machine-name和docker-mashie rm machine-name來停止和移除machines。請注意當使用rm時,是不會提示你是否確定刪除。
$ 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 $
總結,以上就是docker-machie的總覽。正如你看到的,它確實很方便在很多不同的供應商中來引導Docker服務實例,并且使用一個本地machine命令就可以操控他們。
Docker Compose
Docker開始充滿生機是因為有了Fig, 這是我曾在以前的文章中寫到過,當前版本并沒有添加太多的新功能,但它開始奠定了與docker-swarm工作的基礎,單擊 這里查看詳細日志。
像docker-machine一樣,我使用Homebrew公式安裝docker-compose。
$ brew install docker-compose ==> Downloading https://homebrew.bintray.com/bottles/fig-1.1.0.yosemite.bottle.1.tar.gz
############################################################ 100.0%</h6>
==> 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 $ </pre>
然后,使用docker-machine創建一個Docker服務實例:$ docker-machine create --driver virtualbox compose INFO[0001] Creating SSH key... INFO[0001] Creating VirtualBox VM... INFO[0007] Starting VirtualBox VM... INFO[0008] Waiting for VM to start... INFO[0041] "compose" has been created and is now the active machine. INFO[0041] To point your Docker client at it, run this in your shell: $(docker-machine env compose)
因為docker-compose不直接與docker-machine交互,我們需要告訴docker客戶端那些剛剛啟動的服務器實例的詳細信息。
$ $(docker-machine env compose)
此命令注入所需Docker客戶端的環境變量來連接到服務實例,要看到他們,你只需運行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一樣,除了fig.yml文件現在應該改為docker-compose.yml,在我以前的博文里有一個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
它啟用兩個容器并且把它們連接到一起,還有在Nginx容器內的/var/www/html被掛載到host的./web文件夾下。我準備運行docker-compose命令的文件夾的結構是這樣的:
$ tree -a . ├── [russ 356] docker-compose.yml └── [russ 102] web └── [russ 67] index.php1 directory, 2 files</pre>
開始我要拉取需要啟用的鏡像,你可以忽略此部分。$ docker-compose pull Pulling db (russmckendrick/mariadb:latest)... Pulling web (russmckendrick/nginx-php:latest)...
現在鏡像已經被拉取下來,是時候開啟容器了:
$ docker-compose up -d Creating example_db_1... Creating example_web_1...
我們現在有了兩個正在運行的容器:
$ docker-compose psName 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 </pre>
你也可以打開瀏覽器:open http://$(docker-machine ip)
在我的例子中我看到PHPinfo()頁面,如下:
一旦容器開啟,你可以使用docker exec來連接到容器內部:
$ docker exec -it example_web_1 bash [root@997bbe6b5c80 /]# ps aux USER PID %CPU %MEM VSZ RSS TTY STAT START TIME COMMAND root 1 0.2 1.5 115200 15360 ? Ss 13:59 0:01 /usr/bin/python /usr/bin/supervisord -n root 16 0.0 3.2 382876 33624 ? S 13:59 0:00 php-fpm: master process (/etc/php-fpm.conf) root 17 0.0 0.2 110016 2096 ? Ss 13:59 0:00 nginx: master process nginx nginx 18 0.0 0.5 110472 5568 ? S 13:59 0:00 nginx: worker process webserv+ 19 0.0 1.5 383132 16284 ? S 13:59 0:00 php-fpm: pool mywebsite webserv+ 20 0.0 0.8 382876 8848 ? S 13:59 0:00 php-fpm: pool mywebsite webserv+ 21 0.0 0.8 382876 8848 ? S 13:59 0:00 php-fpm: pool mywebsite webserv+ 22 0.0 0.8 382876 8848 ? S 13:59 0:00 php-fpm: pool mywebsite webserv+ 23 0.0 0.8 382876 8852 ? S 13:59 0:00 php-fpm: pool mywebsite root 95 0.0 0.4 91540 4740 ? Ss 13:59 0:00 /usr/libexec/postfix/master -w postfix 97 0.0 0.6 91712 6508 ? S 13:59 0:00 qmgr -l -t unix -u postfix 200 0.0 0.6 91644 6232 ? S 14:05 0:00 pickup -l -t unix -u root 234 2.3 0.2 11748 2968 ? S 14:07 0:00 bash root 250 1.0 1.1 110012 11616 ? S 14:07 0:00 nginx root 251 0.0 0.2 19756 2212 ? R+ 14:07 0:00 ps aux [root@997bbe6b5c80 /]# exit exit $
最后你可以停止以及移除容器,當然還有Docker實例:
$ 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
在進一步討論之前,先看下官方文檔的警告:
警告:Swarm當前是beta版本,因此后期可能會有變化。我們還不推薦在生產環境中使用。現在讓我們使用Homebrew來安裝docker-swarm:
$ brew install docker-swarm ==> Downloading https://homebrew.bintray.com/bottles/docker-swarm-0.1.0.yosemite.bottle.tar.gz
############################################################ 100.0%</h6>
==> Pouring docker-swarm-0.1.0.yosemite.bottle.tar.gz /usr/local/Cellar/docker-swarm/0.1.0: 4 files, 8.7M</pre>
因為我們已經安裝了docker-machine,我將要使用它在本地創建一個集群。首先,我們需要啟動一個實例并運行swarm容器:$ docker-machine ls NAME ACTIVE DRIVER STATE URL SWARM $ docker-machine create -d virtualbox local INFO[0001] Creating SSH key... INFO[0001] Creating VirtualBox VM... INFO[0006] Starting VirtualBox VM... INFO[0006] Waiting for VM to start... INFO[0039] "local" has been created and is now the active machine. INFO[0039] 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 $ </pre>
正如你所見,當容器啟動時,我得到了一個token:63e7a1adb607ce4db056a29b1f5d30cf,我將要用它來添加更多的節點,但是首先我們需要創建一個Swarm master:$ docker-machine create \ → -d virtualbox \ → --swarm \ → --swarm-master \ → --swarm-discovery token://63e7a1adb607ce4db056a29b1f5d30cf \ → swarm-master INFO[0000] Creating SSH key... INFO[0000] Creating VirtualBox VM... INFO[0006] Starting VirtualBox VM... INFO[0006] Waiting for VM to start... INFO[0038] Configuring Swarm... INFO[0043] "swarm-master" has been created and is now the active machine. INFO[0043] To point your Docker client at it, run this in your shell: $(docker-machine env swarm-master)
然后,我們需要連接Docker客戶端到Swarm上,這就需要將--swarm添加到$(docker-machine env machine-name)命令上:
$ $(docker-machine env --swarm swarm-master)
現在讓我們添加另一個節點:
$ docker-machine create \ → -d virtualbox \ → --swarm \ → --swarm-discovery token://63e7a1adb607ce4db056a29b1f5d30cf \ → swarm-node-00 INFO[0000] Creating SSH key... INFO[0000] Creating VirtualBox VM... INFO[0006] Starting VirtualBox VM... INFO[0006] Waiting for VM to start... INFO[0039] Configuring Swarm... INFO[0048] "swarm-node-00" has been created and is now the active machine.
我們現在有了2個節點的集群 - “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來獲取更多有關集群的信息:
$ 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,但這意味著什么?
讓我們拉取一些鏡像:
$ 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鏡像以及在swarm-node-00上拉取mysql的,現在我可以保證容器只在有鏡像的那個節點上啟用:
$ 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
另一個例子是使用節點的端口,讓我們在兩個節點上都拉取我的nginx-php鏡像:
$ docker -H 192.168.99.101:2376 pull russmckendrick/nginx-php $ docker -H 192.168.99.102:2376 pull russmckendrick/nginx-php
現在,讓我們啟用一個容器并綁定到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
你會說這沒什么特別的啊?正常來說,當試圖啟動第二個容器時,你會得到如下信息因為你不用將同一個端口綁定到兩個容器上:
$ docker run -d -p 80:80 russmckendrick/nginx-php FATA[0000] Error response from daemon: unable to find a node with port 80 available
然而,在集群的情況下,因為Docker知道集群節點運行的是什么以及哪些端口是在使用的。Docker可以簡單地通過Swarm在“swarm-node-00”上啟動容器并且它知道“swarm-master”已經使用了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
所有這一切都沒有提示或特殊的命令行,它的幫助僅僅是用它來做到這些。
正如你所看到的,docker-swarm仍然有非常大的發展潛力,但也有一些不如意的地方,如容器不能夠跨節點通訊。然而,伴隨著 socketplane.io(他們使用Open vSwitch開發了一個基于軟件定義網絡解決方案的容器)的加入,我想用不了多長時間這個問題就能得到解決。
最后,讓我們刪除在運行的實例:
$ docker-machine rm local swarm-master swarm-node-00
就這樣吧,期待在未來的幾個月這些工具的更新,我也會進一步跟進。
原文:Docker Machine, Compose & Swarm (翻譯:田浩浩 校對:李穎杰)
來自:http://dockerone.com/article/275