Easy Raspberry PI Cluster Setup with Cloudmesh SDCard Burner

Set up many SD Cards that are preconfigured to create PI clusters.
By Gregor von Laszewski (laszewski@gmail.com, laszewski.github.io), Richard Otten, Anthony Orlowski | 2021.02.07

NOTE: This tutorial has been superseded by new updates. Please see our updated tutorials at:

In this tutorial, we explain how to easily set up a cluster of Pis while burning preconfigured SD Cards. We assume you use an SD Card reader/writer that is plugged into your manager PI that we configure initially with Pi Imager.

Learning Objectives

Topics covered

1. Introduction

Over time we have seen many efforts to create Clusters using Pi’s as a platform. There are many reasons for this. You have full control over the PIs, you use an inexpensive platform, and you use a highly usable platform and provides an enormous benefit to get educated about cluster computing in general.

There are different methods of how to set up a cluster. This includes setups that are known under the terms headless, network booting, and booting from SD Cards. Each of the methods has its advantages and disadvantages. However, the last method is most familiar to the users in the Pi community that come from single Pis. While reviewing the many efforts that describe a cluster set up most of them contain many complex steps that require a significant amount of time as they are executed individually on these Pis. Even starting is non-trivial as a network needs to be set up to access them.

Despite the much improved Pi imager and the availability of Pi bakery, the process is still involved. So we started asking:

Is it possible to develop a tool that is specifically targeted to burn SDCards for the PIs in a cluster one at a time, so we can just plug the cards in, and with minimal effort start the cluster that simply works?

You are in luck, we have spent some time to develop such a tool and present it at as part of PiPlanet1 . No more spending hours upon hours to replicate the steps, learn complex DevOps tutorials, but instead get a cluster set up easily with just a few commands.

For this, we developed cms burn which is a program that you can execute either on a “manager” Pi (or in a Linux or macOS computers) to burn cards for your cluster.

We have set up on GitHub a comprehensive package that can be installed easily we hope that it is useful to you. All of this is discussed in detail at the cloudmesh-pi-burn README2. There you can also find detailed instructions on how to use a Mac or Linux computer to burn directly from them. To showcase to you how easy it is to use we demonstrate here the setup of a cluster with five nodes.

2. Requirements

For a list of possible part choices, please see:

3. The Goal

We will be creating the following setup using 5 Raspberry Pis (you need a minimum of 2, but our method works also for larger numbers of PIs). Consequentially, you will also need 5 SD cards for each of the 5 Pis. You will also want a network switch (managed or unmanaged) with 5 ethernet cables (one for each Pi).

Figure 1 shows our network configuration. From the five Raspberry Pis, one is dedicated as a manager and four as workers. We use WiFi between the manager PI to allow for you to set it up anywhere in your house or dorm (other configurations are discussed in the README).

We use an unmanaged network switch, where the manager and workers can communicate locally with each other, and the manager provides internet access to the workers via a bridge that we configure for you.

Figure 1: Pi Cluster setup with bridge network

4. Steps

Step 1. Burning and Configuring the Manager

Choose one SD card for the manager (yellow card in Figure 1). Using your laptop, download the Raspberry Pi Imager RaspberryOS desktop version. This is the recommended version for PIs at this time. We use Pi Imager to burn our manager. Note this is the only time we will need to use PI Imager.


Figure 2. Pi Imager

Write to your SD card. Once the process is complete and verified, insert the SD Card into your manager Pi. Connect your manager to the peripherals (keyboard, mouse, monitor, power). Switch on the power.

Note you may also use a headless setup. See here for more information on headless setups.

Next, we walk through the initial setup process of the Pi and configure the settings in accordance with your situation. We have provided screenshots in Figures 3 - 8 that depicts this process.


Figure 3. After successfully booting the Pi will display the Welcome Page.


Figure 4. Set country, language, and timezone. Additionally, you must use the proper keyboard layout. For the US enable “Use US Keyboard”.


Figure 5. Set your password and use a strong password.


Figure 6. Choose your Wifi network and configure it while adding the password.


Figure 7. The setup prompt will ask you if you wish to update the software. You may do so, or you may skip, as the installation script that we will run will do this for you.


Figure 8. Setup is now complete.

Step 2. Create an SSH key

As we will be using keys to authenticate with the workers, you need to create one in a terminal with

pi@managerpi:~ $ ssh-keygen

It will ask you for a location, choose the default. It will also ask you for a passphrase. Please use a strong one and do not make it the same as the password for your manager PI.

Step 3. Installing Cloudmesh

Now let us install cloudmesh burn, which allows us to burn preconfigured SD Cards for clusters easily. Open a new terminal window and run the following command. To make the installation and needed updates to your PI simple, we have provided a one-line install script that you can run via curl:

pi@managerpi:~ $ curl -Ls http://cloudmesh.github.io/get/pi | sh -

This will set up a python venv on your computer manager Pi. It may take 5-7 minutes as it will also update your Pi and install all other requirements.

You will want to reboot your Pi after this.

pi@managerpi:~ $ sudo reboot

Step 4. Creating our Cluster Inventory

To manage information about our cluster, we will use a Cloudmesh Inventory. This will allow you to easily track and manage the configuration of your cluster nodes. Let us create an inventory for our cluster as follows:

(ENV3) pi@managerpi:~ $ cms inventory create --hostnames="managerpi,worker00[1-4]" --ip="10.1.1.[1-5]"  --inventory=cluster.yaml latest-lite

You can inspect the inventory with the list command as shown next. Double-check if it looks like:

(ENV3) pi@managerpi:~ $ cms inventory list --inventory=cluster.yaml

+-----------+-----------+------+-------------+---------+-------+---------+----------+----------+-----+---------+--------+---------+-------------+-------------------+----------+
| host      | name      | type | tag         | cluster | label | service | services | ip       | dns | project | owners | comment | description | keyfile           | status   |
+-----------+-----------+------+-------------+---------+-------+---------+----------+----------+-----+---------+--------+---------+-------------+-------------------+----------+
| managerpi | managerpi |      | latest-lite | cluster |       | manager |          | 10.1.1.1 |     |         |        |         |             | ~/.ssh/id_rsa.pub | inactive |
| worker001 | worker001 |      | latest-lite | cluster |       | worker  |          | 10.1.1.2 |     |         |        |         |             | ~/.ssh/id_rsa.pub | inactive |
| worker002 | worker002 |      | latest-lite | cluster |       | worker  |          | 10.1.1.3 |     |         |        |         |             | ~/.ssh/id_rsa.pub | inactive |
| worker003 | worker003 |      | latest-lite | cluster |       | worker  |          | 10.1.1.4 |     |         |        |         |             | ~/.ssh/id_rsa.pub | inactive |
| worker004 | worker004 |      | latest-lite | cluster |       | worker  |          | 10.1.1.5 |     |         |        |         |             | ~/.ssh/id_rsa.pub | inactive |
+-----------+-----------+------+-------------+---------+-------+---------+----------+----------+-----+---------+--------+---------+-------------+-------------------+----------+

Step 5. Burning the SD Cards

We can now begin burning.

You can now plug in your SD Card reader/writer into the managerpi. Ensure you have also inserted an SD card into your reader/writer. Warning this SD Card will be formatted, thus all content will be deleted and lost.

Verify your device is detected with the following command:

(ENV3) pi@managerpi:~ $ cms burn info

# ----------------------------------------------------------------------
# SD Cards Found
# ----------------------------------------------------------------------

+----------+------------------------+-------------+------------------+--------------+------------+---------+----------+-------------+-------------+
| Path     | Info                   | Formatted   | Size             | Plugged-in   | Readable   | Empty   | Access   | Removable   | Writeable   |
|----------+------------------------+-------------+------------------+--------------+------------+---------+----------+-------------+-------------|
| /dev/sdb | Generic STORAGE DEVICE | True        | 64.1 GB/59.7 GiB | True         | True       | False   | True     | True        | True        |
+----------+------------------------+-------------+------------------+--------------+------------+---------+----------+-------------+-------------+
Note we omit some information from cms burn info for simplicity

From cms burn info, we see our device is /dev/sdb. Note this may be different on your Pi. If your device is not showing up, ensure you have an SD Card inserted, and try unplugging and plugging the SD Card reader/writer.

We can now begin burning our cluster. The following command will download the necessary Raspberry Pi OS images, configure manager as a Wifi bridge to provide internet access to workers and burn the SD Cards. Note you will need to cycle SD cards after each burn.

(ENV3) pi@managerpi:~ $ cms burn create --inventory=cluster.yaml --device=/dev/sdb --name=managerpi,worker00[1-4]

Manager hostname is the same as this system's hostname. Is this intended? (Y/n) Y
Do you wish to configure this system as a WiFi bridge? A restart is required after this command terminates (Y/n) Y
Some output of cms burn has been omitted for simplicity. Note that image extraction may take more than a minute.

As each SD Card is burned, cms burn will prompt you to insert a new SD Card to be burned.

After all the cards are burned, plug them into your worker Pis and boot. Reboot the managerpi.

(ENV3) pi@managerpi:~ $ sudo reboot

Step 6. Verifying the Workers

Once your workers are booted, you can verify the connection with the following simple command. This command will return the temperature of the Pis.

(ENV3) pi@managerpi:~ $ cms pi temp worker00[1-4]
pi temp worker00[1-4]
+-----------+--------+-------+----------------------------+
| host      |    cpu |   gpu | date                       |
|-----------+--------+-------+----------------------------|
| worker001 | 36.511 |  36.5 | 2021-02-22 00:06:48.873427 |
| worker002 | 36.998 |  37   | 2021-02-22 00:06:48.813539 |
| worker003 | 36.998 |  37   | 2021-02-22 00:06:48.843944 |
| worker004 | 36.498 |  36   | 2021-02-22 00:06:48.843956 |
+-----------+--------+-------+----------------------------+

5. Using the Pis

As we use ssh keys to authenticate between manager and workers, you can directly log into the workers from the manager.

More details are provided on our web pages at

Other cloudmesh components are discussed in the cloudmesh manual4.

Acknowledgement

We would like to thank the following community members for testing the recent versions: Venkata Sai Dhakshesh Kolli, Rama Asuri, Adam Ratzman. Previous versions of the software obtained code contributions from Sub Raizada, Jonathan Branam, Fugang Wang, Anand Sriramulu, Akshay Kowshik.

References

  1. PiPlanet Web Site, https://piplanet.org ↩︎
  2. Cloudmesh pi burn README, https://github.com/cloudmesh/cloudmesh-pi-burn/blob/main/README.md ↩︎
  3. Parts for building clusters, https://cloudmesh.github.io/pi/docs/hardware/parts/ ↩︎
  4. Cloudmesh Manual, https://cloudmesh.github.io/cloudmesh-manual/ ↩︎