Embedded Developer

A common problem in the IT world is to create a consistent environment in a repeatable manner. This is important in a number of use cases:

• Development
• Testing
• Training

This series of posts will investigate using Ansible to create a consistent test environment, one that can be setup and torn down quickly and easily.

The starting point is setting up the hardware and installing the operating system (OS) which will be covered here. Subsequent posts will use Ansible to configure the system and deploy additional tools.

The Hardware

The test environment will be based around the Raspberry Pi 5 (although any version of the Pi hardware could be used). The system will be built around the following components:

• Raspberry Pi (3, 4 or 5)
• 256 GByte SATA SSD
• SATA to USB adapter
• Cooling fan (for the Raspberry Pi 5)
• Power Supply
• Ethernet cable
• 3D printed mounts to bring everything together

Grabbing a Raspberry Pi 5 and putting all of this together yields something like this:

Raspberry Pi Setup

SATA SSDs have been chosen for the OS and data storage as they are both faster and more reliable than SD cards. From a cost perspective they are not too much more expensive than a quality SD card. It should be noted that recent third party addon boards are becoming available that add one or two NVMe drives to be added to the the Raspberry Pi 5 using the PCIe bus.

Write OS Image

The easiest way to create a bootable Raspberry Pi system is to use the Raspberry Pi Imager. This is a free tool that allows the selection of one of the many operating systems available for the Raspberry Pi and it can then be used to write the operating system to a SD card or HDD/SSD

The process starts by connecting the SATA to USB adapter the the SSD and then connecting the drive to the host computer. This makes the drive appear as an external USB drive.

Now start Raspberry Pi Imager:

Raspberry Pi Imager

Select the device we are going to create the image for, in this case this is the Raspberry Pi 5:

Select Device

The next step is to decide which operating system should be installed on the SSD. There are a large number of options and the selection will depend upon what you want to achieve. In this case we can use a basic system such as Raspberry Pi OS Lite. Firstly, select the Raspberry Pi (64-bit) operating system:

Select Operating System

Now refine this selection and select the Raspberry Pi OS Lite (64-bit):

Select Raspberry Pi Lite

A basic system will be adequate as the device is intended to be run headless and so the desktop environment and applications are not required.

Next step is to select the storage device that the image will be written to. Once this is done we can move on to providing some configuration options for the operating system.

Ready For Configuration

Click the Next button to move on to the next step, editing the configuration.

Edit Settings

Clicking Edit setting starts the editing process. The General options are presented first, here we can set the following:

• Hostname
• User name and password
• WiFi access point details

Customise General Settings

SSH should be enabled in order to run the system headless. This is enabled on the Services tab:

Customise Services

Clicking on Save now gives the option of applying the settings and start writing the image to the SSD:

Apply Settings

The final step is to verify that the SSD can be erased:

Confirm Media Erase

Control now passes back to the main window where the write and verification progress can be monitored:

Writing OS

After a short while the the process will complete and Raspberry Pi Imager wil conform that the image has been written successfully and the drive can now be disconnected from the host computer and connected to the Raspberry Pi 5:

OS Write Successful

Conclusion

The whole process of creating the image is straightforward and only takes a few minutes. At the end of the process the Raspberry Pi is ready to boot.

The next step will be to start the installation and configuration of additional software tools and components. Something for the next post in this series.

Recent work has involved reviewing some test environments for an IoT development board. The aim is to improve some of the components used for testing as well as adding new functionality. The requirements are:

• Provide an updated version of existing functionality
• Single board environment with all functionality deployed for quick testing
• Cluster distributing the test environment for load testing

The most cost effective way to do this is to use a number of Raspberry Pi single board computers. These boards are now becoming available in quantities after several years of limited availability.

The Problem

How to setup the environment in such a way that will allow a fresh environment to be created reliably.

Enter ansible.

Ping

First step, try to contact a board and this is where ping comes in. This command will verify that ansible can connect to a board. The following command will test the connection to each board:

ansible cluster -m ping -i hosts

This command requires a text file hosts containing the list of boards to the contacted. The file is simple and may only contact two lines:

[cluster]
node

In the above example, the file defines a group of machines to be contacted and this is named cluster and in this case the group contains only one machine and this is named node. The name cluster is also mentioned in the ansible command above.

Additional machines can also be named under the cluster entry by simply placing additional entries on a new line in the file.

So far this is nothing new and it is covered in the Ansible documentation.

What Happened

The first step was to use the Raspberry Pi Imager application to create a new image on a new SSD. Nothing complex:

• Raspberry Pi 64-bit Lite OS
• Set the machine name to be node
• Enable SSH
• Set the user name to clusteruser and give the user a secure password

The password was then stored on the local machine in an environment variable CLUSTER_PASSWORD to allow the scripts to be stored in source control without giving away any secrets.

Time to test the connection with the following command:

ansible cluster -m ping -i hosts --extra-vars "ansible_user=clusteruser ansible_password=$CLUSTER_PASSWORD"

Breaking this down, we want to ping all of the machines defined in the cluster group. The group is defined in the file hosts and we are going to log on to the machines with the user name clusteruser and with the password contained in the CLUSTER_PASSWORD environment variable.

Now running the above command results in the following:

node | SUCCESS => {
    "ansible_facts": {
        "discovered_interpreter_python": "/usr/bin/python3"
    },
    "changed": false,
    "ping": "pong"
}

Conclusion

A good start to the project, now on to something more complex, time to install and configure some software.

And I can’t believe I’ve missed Ansible for so long.