Alexander Beletsky's development blog

My profession is engineering

Playing with Dokku on Vagrant

As I said previously, it’s very easy to turn Linux machine into Heroku-like server. But, before setting up paying account on Amazon or Digital Ocean, it’s nice to just play it locally. Will do that, running a Dokku on virtual machine. Will setup development environment and do first local deployment, just to see some real features.

It does not require complex environment to run Dokku locally. All you need is git and Vagrant.

Prepearing Environment

You box should contain few things: git (github account), VirtualBox and Vagrant. If you don’t have Vagrant installed, please do now. It makes a lot of sense to keep such kind software on machine.

Here you can find some instructionn of how to do that.

Cloning Dokku

You should clone Dokku repo locally. In your development folder, run

1
2
> git clone git@github.com:progrium/dokku.git
> cd dokku

Dokku repository already contains Vagrant file, with all required configuration.

Local Networking

Just in sake of convenience, will map 10.0.0.2 ip address (the one that Vagrant machine is assigned with), to dokku.me DNS name, so it’s just handy for testing.

1
› sudo nano /private/etc/hosts

and put last 2 lines, as it shown in my example:

1
2
3
4
5
6
7
8
9
10
11
12
##
# Host Database
#
# localhost is used to configure the loopback interface
# when the system is booting.  Do not change this entry.
##
127.0.0.1       localhost
255.255.255.255 broadcasthost
::1             localhost
fe80::1%lo0     localhost
10.0.0.2        dokku.me
10.0.0.2        node-simple.dokku.me

Fire Up Virtual Machine

In dokku folder, you should run

1
> vagrant up

It will start to prepare new virtual environment for you,

1
2
3
4
5
6
7
8
9
10
11
12
13
 vagrant up
Bringing machine 'default' up with 'virtualbox' provider...
[default] Setting the name of the VM...
[default] Clearing any previously set forwarded ports...
[default] Creating shared folders metadata...
[default] Clearing any previously set network interfaces...
[default] Preparing network interfaces based on configuration...
[default] Forwarding ports...
[default] -- 22 => 2222 (adapter 1)
[default] -- 80 => 8080 (adapter 1)
[default] Running any VM customizations...
[default] Booting VM...
[default] Waiting for VM to boot. This can take a few minutes.

I have a problem with this step few times, so if you machine could not be booted, try to run vagrant reload, it should help.

It would took about up to 20 minutes, to fire up virtual machine, install git there, install Docker + Dokku. As soon as it’s done, it’s possible to access machine by ssh.

Last thing you need to do, is to upload your ssh key to Dokku server, so you will be to git push code there.

1
> cat ~/.ssh/id_rsa.pub | ssh root@dokku.me "sudo gitreceive upload-key alexanderbeletsky"

You can use default vagrant root password, which is vagrant.

Now, just to check that everything is fine simply access you machine by ssh,

1
> vagrant ssh

So, just try that everything is running fine by checking version of Docker:

1
2
> docker -v
Docker version 0.6.1, build 5105263

The instance is ready for deployment.

Deploy to Dokku

If you still there, you can just quit vagrant ssh session, and went to the folder with Node.js application. I’ll be using really simple app, called Node-simple – express.js based, that serves one HTML file and shows NODE_ENV env. variable.

So, what you need to setup is remote repository to push to:

1
> git remote add local-deploy git@dokku.me:node-simple

That’s all. You ready for first deploy, just push code to machine with Dokku:

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
git push local-deploy master
› git push local-deploy master --force
Counting objects: 5, done.
Delta compression using up to 8 threads.
Compressing objects: 100% (3/3), done.
Writing objects: 100% (3/3), 289 bytes | 0 bytes/s, done.
Total 3 (delta 2), reused 0 (delta 0)
-----> Building node-simple ...
       Node.js app detected
-----> Resolving engine versions
       Using Node.js version: 0.8.25
       Using npm version: 1.2.30
-----> Fetching Node.js binaries
-----> Vendoring node into slug
-----> Installing dependencies with npm
       npm WARN package.json application-name@0.0.5 No repository field.
       npm WARN package.json application-name@0.0.5 No readme data.
       npm WARN package.json send@0.1.0 No repository field.
       ...
=====> Application deployed:
       http://node-simple.dokku.me

To git@dokku.me:node-simple
   dd05aae..ac5b6da  master -> master

Setting up Environment

Every application requires environment. It’s common practice to set at least NODE_ENV variable for Node.js applications. To do that, you need to create ENV file inside the /home/git/node-simple folder.

1
› ssh root@dokku.me "echo export NODE_ENV="development" > /home/git/node-simple/ENV"

Now, let’s re-deploy the application, change the version in package.json and push code again.

1
› git push local-deploy master

Now, application is ready to be accessed.

Accessing the Application

Open Chrome and hit http://node-simple.dokku.me, so you will see such response:

You can play a bit further, by just changing some Node.js code of end-points and re-deploy. Each time, new docker instance is started and served on http://node-simple.dokku.me. The experience of deployment is really like something you have with Heroku.

Just looking on logs while new application is deployed, would give pretty much idea of what’s going on there.

So, your local Vagrant image will good start place, before you ready to use Dokku on cloud.