Skip to main content

raspberrypi3

Getting Started on the Raspberry Pi 3 B/B+

In balenaOS all application logic and services are encapsulated in Docker containers. In this getting started guide we will walk you through setting up one of our pre-built development OS images and creating a simple application container. In the guide we will use the balena CLI tool to make things super easy.

Download an Image​

To get a balenaOS device setup, we will first need to flash a system image on to the device, so head over to balena.io/os and grab the development OS for your board.

Once the download is finished, make sure to unzip the file and keep the resulting balena.img somewhere safe, we will need it very soon!

Install the Balena CLI​

The CLI is a collection of utilities which helps us to develop balenaOS based application containers.

Install the CLI with the following instructions, available here.

Configure the Image​

To allow balenaOS images to be easily configurable before boot, some key config files are added to boot partition. In this step we will use the CLI to configure the network, set our hostname to mydevice and disable persistent logging, because we don’t want to kill our flash storage with excessive writes. If you are using ethernet you can skip this step, by default the device hostname will be balena.

Note: Check our list of supported WiFi adapters and modems.

$ sudo balena local configure ~/Downloads/balena.img
? Network SSID I_Love_Unicorns
? Network Key superSecretPassword
? Do you want to set advanced settings? Yes
? Device Hostname mydevice
? Do you want to enable persistent logging? no
Done!

Get the Device Up and Running​

Okay, so now we have a fully configured image ready to go, so let’s flash and boot this baby. For this step we recommend balenaEtcher a handy flashing utility. You can, however, flash this image using dd or any other SD card flashing utility, if you wish.

Flash SD card

Open balenaEtcher and use the blue "Select Image" button to find the disk image you downloaded earlier. Once you have selected the image you want to flash, insert your SD card or connect your device (in the case of a balenaFin) to your laptop and flash the image.

Boot the device​

Now power on and boot up your device, after about 10 seconds or so your device should be up and connected to your local network, you should see it broadcasting itself as mydevice.local. To check this, let’s try ping the device.

$ ping mydevice.local
PING 192.168.1.111 (192.168.1.111): 56 data bytes
64 bytes from 192.168.1.111: icmp_seq=0 ttl=64 time=103.674 ms
64 bytes from 192.168.1.111: icmp_seq=1 ttl=64 time=9.723 ms
^C
--- 192.168.1.111 ping statistics ---
6 packets transmitted, 6 packets received, 0.0% packet loss
round-trip min/avg/max/stddev = 7.378/24.032/103.674/35.626 ms

Now if we want to poke around a bit inside balenaOS we can just ssh in with:

$ balena ssh mydevice.local
Last login: Wed Sep 11 10:32:01 2019 from fe80::1420:6988:2a39:3d3d%wlan0
root@mydevice:~# uname -a
Linux mydevice 4.19.66 #1 SMP PREEMPT Wed Aug 28 16:57:11 UTC 2019 aarch64 aarch64 aarch64 GNU/Linux
root@mydevice:~# balena-engine version
Client:
 Version:           18.09.8-dev
 API version:       1.39
 Go version:        go1.10.8
 Git commit:        80d443d400dcc85e87322f72866593b46bafb157
 Built:             Mon Aug 26 09:22:28 2019
 OS/Arch:           linux/arm64
 Experimental:      false

Server:
 Engine:
  Version:          18.09.8-dev
  API version:      1.39 (minimum version 1.12)
  Go version:       go1.10.8
  Git commit:       80d443d400dcc85e87322f72866593b46bafb157
  Built:            Mon Aug 26 09:22:28 2019
  OS/Arch:          linux/arm64
  Experimental:     true

Running your first Container​

Clone a demo Project​

$ git clone https://github.com/balenalabs/multicontainer-getting-started.git

Get some Containers Running​

To launch containers on our device, we will use the balena push command. First we need to change directory into the project we cloned down and then issue the following command.

$ balena push mydevice.local

This command will use the image specified by the docker-compose.yml or Dockerfile in the root of your project directory. The build of this project will happen on your balenaOS device and once completed, the command will start up a container from that newly built image(s). Note: The first time you push a project it will need to pull the base images, so it can take a few minutes. Later pushes will be much faster.

Once it's finished building, you should see [Logs] ... streaming back and if you visit http://mydevice.local/ you will get a pretty little webpage showing CPU load and memory usage. You will also notice that anytime you change the code in the multicontainer-getting-started code base it will hot reload your containers, syncing the new code changes on the device.

Poking Around balenaOS​

To help explore balenaOS devices and application containers more easily, the balena CLI has an ssh command which will help you connect either to the HostOS or a running container on the device.

To ssh into the hostOS:​

$ balena ssh mydevice.local

To ssh into a particular container:​

$ balena ssh mydevice.local <service_name>

Where <service_name> can be replaced with any of the containers defined in your docker-compose.yml. If you have a single container app that doesn't have a compose file, your service will get the default name of main, so you can access the container using balena ssh mydevice.local main.

Going Further​

Advanced Settings​

Either mount the SD card and run: 
$ sudo balena local configure /path/to/drive

And select y when asked if you want to add advanced settings.

Alternatively you can add “persistentLogging”: true to config.json in your boot partition of the SD card

To enable persistent logs in a running device, add “persistentLogging”: true to /mnt/boot/config.json and reboot.

The journal can be found at /var/log/journal/ which is bind mounted to root-overlay/var/log/journal in the resin-conf partition. When logging is not persistent, the logs can be found at /run/log/journal/ and this log is volatile so you will lose all logs when you power the device down.

Creating a Project from Scratch​

Alright! So we have an awesome container machine up and running on our network. So let’s start pushing some application containers onto it. In this section we will do a quick walk through of setting up a Dockerfile and make a simple little node.js webserver.

To get started, let’s create a new project directory called “myapp” and create a new file called Dockerfile.

$ mkdir -p myapp && touch Dockerfile

Now we will create a minimal node.js container based on the slim Alpine Linux distro. We do this by adding the following lines to our Dockerfile.

FROM balenalib/raspberrypi3-alpine-node

CMD ["cat", "/etc/os-release"]

The FROM tells Docker what our container will be based on. In this case an Alpine Linux userspace with just the bare essentials needed for the node.js runtime. The CMD just defines what our container runs on startup. In this case, it’s not very exciting yet.

Now to get our application running on our device we can use the balena push mydevice.local functionality.

$ balena push mydevice.local

[Info]    Starting build on device mydevice.local

[Info]    Creating default composition with source: .

[Build]   [main] Step 1/4 : FROM balenalib/raspberrypi3-alpine-node

[Build]   [main]  ---> 25d3c8c8d0af

[Build]   [main] Step 2/4 : CMD ["cat", "/etc/os-release"]

[Build]   [main]  ---> Running in 449e2d11b48f

[Build]   [main] Removing intermediate container 449e2d11b48f

[Build]   [main]  ---> fc19268cf077

[Build]   [main] Step 3/4 : LABEL io.resin.local.image=1

[Build]   [main]  ---> Running in efbf5e0d1027

[Build]   [main] Removing intermediate container efbf5e0d1027

[Build]   [main]  ---> da6d888c128f

[Build]   [main] Step 4/4 : LABEL io.resin.local.service=main

[Build]   [main]  ---> Running in 4befab0ab3f3

[Build]   [main] Removing intermediate container 4befab0ab3f3

[Build]   [main]  ---> 8351b83ebb73

[Build]   [main] Successfully built 8351b83ebb73

[Build]   [main] Successfully tagged local_image_main:latest

[Live]    Waiting for device state to settle...

[Info]    Streaming device logs...

[Live]    Watching for file changes...

[Logs]    [9/11/2019, 12:50:23 PM] Creating volume 'resin-data'

[Logs]    [9/11/2019, 12:50:26 PM] Installing service 'main sha256:8351b83ebb73f49ae044237191307c3bfed80c2ace2535bf33e8e4361c43461d'

[Logs]    [9/11/2019, 12:50:26 PM] Installed service 'main sha256:8351b83ebb73f49ae044237191307c3bfed80c2ace2535bf33e8e4361c43461d'

[Logs]    [9/11/2019, 12:50:26 PM] Starting service 'main sha256:8351b83ebb73f49ae044237191307c3bfed80c2ace2535bf33e8e4361c43461d'

[Logs]    [9/11/2019, 12:50:27 PM] Started service 'main sha256:8351b83ebb73f49ae044237191307c3bfed80c2ace2535bf33e8e4361c43461d'

[Logs]    [9/11/2019, 12:50:27 PM] [main] NAME="Alpine Linux"

[Logs]    [9/11/2019, 12:50:27 PM] [main] ID=alpine

[Logs]    [9/11/2019, 12:50:27 PM] [main] VERSION_ID=3.10.2

[Logs]    [9/11/2019, 12:50:27 PM] [main] PRETTY_NAME="Alpine Linux v3.10"

[Logs]    [9/11/2019, 12:50:27 PM] [main] HOME_URL="https://alpinelinux.org/"{'\n'}
[Logs]    [9/11/2019, 12:50:27 PM] [main] BUG_REPORT_URL="https://bugs.alpinelinux.org/"{'\n'}

This command will start the build on your local balenaOS device from whatever you have in the current working directory. It will then start up all the containers and stream back the logs from each container to the terminal. We can see that for our local balenaOS device we have an app called [main] which will be created from an image called local_image_main and is associated to a container on our device called main_1_1. You will notice that the container keeps restarting over and over. This is due to the fact that the our main process of printing out the os-release file exits after running and by default our containers restart policy is to always restart containers.

So now that we are building, let’s start adding some actual code! We will just add main.js file in the root of our myapp directory.

//main.js
console.log("Hey… I’m a node.js app running in a container!!");

We then make sure our Dockerfile copies this source file into our container context by replacing our current CMD ["cat","/etc/os-release"] in our Dockerfile with the following.

FROM balenalib/raspberrypi3-alpine-node

WORKDIR /usr/src/app

COPY . .

CMD ["node", "main.js"]

This puts all the contents of our myapp directory into /usr/src/app in our running container and says we should start main.js when the container starts.

Alright, so we have a simple javascript container, but that’s pretty boring, let’s add some dependencies and complexity. To add dependencies in node.js we need a package.json, the easiest way to whip up one is to just run npm init in the root of our myapp directory. After a nice little interactive dialog we have the following package.json in directory.

{
  "name": "myapp",
  "version": "1.0.0",
  "description": "a simple hello world webserver",
  "main": "main.js",
  "scripts": {
    "test": "echo \"no tests yet\""
  },
  "repository": {
    "type": "git",
    "url": "none"
  },
  "author": "Shaun Mulligan <[email protected]>",
  "license": "ISC"
}

Now it’s time to add some dependencies. For our little webserver, we will use the popular expressjs module. We can add it to the package.json after the "license": "ISC", so it now looks like this:

{
  "name": "myapp",
  "version": "1.0.0",
  "description": "a simple hello world webserver",
  "main": "main.js",
  "scripts": {
    "test": "echo \"no tests yet\""
  },
  "repository": {
    "type": "git",
    "url": "none"
  },
  "author": "Shaun Mulligan <[email protected]>",
  "license": "ISC",
  "dependencies": {
    "express": "^4.14.0"
  }
}

Now all we need to do is add a few more lines of javascript to our main.js and we are off to the races.

//main.js

var express = require('express');
var app = express();

// reply to request with "Hello World!"
app.get('/', function (req, res) {
  res.send("Hello World, I'm a container running on balenaOS!");
});

//start a server on port 80 and log its start to our console
var server = app.listen(80, function () {

  var port = server.address().port;
  console.log("Hey… I’m a node.js server running in a container and listening on port: ", port);
});

Great, so now we are almost ready to go, but we want to make sure our dependency gets installed when we build. We then need to run a npm install in our build, so we add a few lines to our Dockerfile.

FROM balenalib/raspberrypi3-alpine-node

WORKDIR /usr/src/app

COPY package.json package.json

RUN npm install

COPY . .

CMD ["node", "main.js"]

NOTE: Add node_modules to your .dockerignore file, otherwise your local modules might be copied to the device with the above Dockerfile, and they are likely the wrong architecture for your application!

We can now deploy our new webserver container again with:

$ balena push <DEVICE_IP>

You should now be able to point your web browser on your laptop to the IP address of your device and see the "Hello, World!" message.

Configure balenaOS without the CLI​

If you are not using the CLI, you will need to mount the boot partition of the image and edit the configuration manually.

Edit /boot/config.json so it looks like this:

{
  "persistentLogging": false,
  "hostname": "mydevice",
}

And create a file in /boot/system-connections called my-wifi with the following content and the ssid and psk values replaced as needed.

[connection]
id=my-wifi
type=wifi

[wifi]
mode=infrastructure
ssid=I_Love_Unicorns

[wifi-security]
auth-alg=open
key-mgmt=wpa-psk
psk=superSecretPassword

[ipv4]
method=auto

[ipv6]
addr-gen-mode=stable-privacy
method=auto