Skip to content

Latest commit

 

History

History
93 lines (60 loc) · 8.28 KB

README.md

File metadata and controls

93 lines (60 loc) · 8.28 KB

ALARM (ArchLinuxArm) Imager

I wrote this script as I install Arch Arm Linux (ALARM) on a number of my SBCs as part of a low(er) cost kubernetes cluster I am building at home, and it was a pain to go through multiple devices manually using the Arch Linux Arm website instructions for the board. So far I have implemented a build method for the following Single Board Computers (SBCs) in the script:

  • Banana Pi - The script recognises this board as BANANA_PI. As this is not officially supported by ALARM and I had the most deviations from the guide to get working, this is the most likely to break again.
  • Banana Pro - The script recognises this board as BANANA_PRO. Same as the banana pi in terms of issues.
  • Raspberry Pi model 2B - The script recognises this board as RASPBERRY_PI_3. The script goes as described on the website, except the boot partition is mounted in the root partition to avoid the copy stage.
  • Raspberry Pi model 3B/3B+ - The script recognises this board as RASPBERRY_PI_3. The script goes as described on the website, except the boot partition is mounted in the root partition to avoid the copy stage.
  • Raspberry Pi model 4 - The script recognises this board as RASPBERRY_PI_4. The script goes as described on the website, except the boot partition is mounted in the root partition to avoid the copy stage.
  • Rock 64 - The script recognises this board as ROCK64. The script goes as described on the website.

The automation has three methods of use for creating an image:

  1. A script that writes directly to the SD Card/USB Drive. Be careful you get the correct device (e.g. /dev/sda or /dev/mmcblk0.)
  2. A script that creates an image that can be used to make the above. This is the default method for the script.
  3. A docker image that runs the script. This can be built with the dockerfile or used directly from Docker Hub. They all execute the same process, so use your favourite method.

If you make an image, you can then write it with your favourite image writing solution. Using WSL2 it is possible to build an image from Windows and image it using Rufus or Balena, etc.

Getting Started

The first decision you need to make is how to use the script. I prefer the use of docker, as there is less consideration for dependencies. The use of both the script and docker is described below. I have used the docker successfully with a WSL2 image of Ubuntu as the docker host.

Script

If you want to use the script directly you should first decide if you are going to make an image, or prepare the media (such as sd card) directly. If you are going to image the SD Card directly find the name of the device. I use the following command on Ubuntu and Arch:

lsblk

This should show you the list of block devices connected including the device you want. Typically it will be a '/dev/sdX' or '/dev/mmcblkX', but becareful not to specify the wrong drive.

You can either edit the default variables at the top of the script, or you can use flags to set them. The command with flags looks like:

alarm-imager.sh -m [Image] -d [Working Directory] -b [board] -a [Architecture] -s [Image Size] -h [Help]

The flags that can be set are:

Flags Effect
-m The '-m' flag sets if an image (input Image) or a drive is to used. If it is a drive put the /dev path such as '/dev/mmcblk0', '/dev/sda', 'etc'.
-d The '-d' flag sets if a custom working directory for the script is used. The default location is '/tmp/alarm-image'.
-b The '-b' flag sets the board to build for. The current list of options for this are:
* All Boards = All - Create all possible images with all Arch combinations. This will ignore Mode and only work in Image mode.
* Banana pi = BANANA_PI - arch = armv7 - arch flag will be ignored
* Banana pi pro = BANANA_PRO - arch = armv7 - arch flag will be ignored
* Raspberry pi 2 = RASPBERRY_PI_2 - arch = armv7 - arch flag will be ignored
* Raspberry pi 3 = RASPBERRY_PI_3 - arch = armv7 or arm64
* Raspberry pi 4 = RASPBERRY_PI_4 - arch = armv7 or arm64
* Rock 64 = ROCK64 - arch = arm64 - arch flag will be ignored
The Raspberry Pi 4 board is the default board the script will make an image for.
-a The '-a' flag sets the architecture to build for. This will default to ARM64 when there is a choice between ARM (armv7) and ARM64.
-s The '-s' flag is used to make a custom sized image. The units are MB and the default is 8000. 8000 MB is the minimum size recommended for ALARM images, so the smallest that can be selected.
-n The '-n' flag will remove all items used to create the image. This will also delete any previously cached data.
-h The '-h' flag will list how to use the script and will not proceed to make an image if used.

If the media was manupulated directly, it will be unmounted and ready to use. Otherwise the image will be saved in the Images directory of the working directory in the format ALARM-BOARD-ARCH e.g. ALARM-BANANA_PI-ARM.img. Most of the files used will be saved in the working directory to save time if a new image is made. You can delete these if this is not wanted.

Docker

The docker image can be built from the dockerfile or downloaded directly from github.

To build the image first download the repo, and run the following command from the directory:

docker build . -t bowseruk/alarm-imager:latest

Or to pull it from Docker Hub:

docker pull bowseruk/alarm-imager

Then run the docker run command

docker run -e BOARD=board -e ARCH=architecture -e IMAGE_SIZE=desired_size  -v "/your/directory":/data/alarm-imager --privileged bowseruk/alarm-imager:latest

In this command the environment variables can be set:

Variable Description
BOARD This sets the '-b' flag and chooses the board to build for. The current list of options for this are:
* All Boards = All - Create all possible images with all Arch combinations.
* Banana pi = BANANA_PI - arch = armv7 - arch flag will be ignored
* Banana pi pro = BANANA_PRO - arch = armv7 - arch flag will be ignored
* Raspberry pi 2 = RASPBERRY_PI_2 - arch = armv7 - arch flag will be ignored
* Raspberry pi 3 = RASPBERRY_PI_3 - arch = armv7 or arm64
* Raspberry pi 4 = RASPBERRY_PI_4 - arch = armv7 or arm64
* Rock 64 = ROCK64 - arch = arm64 - arch flag will be ignored
The Raspberry Pi 4 board is the default board the script will make an image for.
ARCH This sets the -a flag which chooses the architecture to build for when there is a choice. This will default to ARM64 when there is a choice between ARM (armv7) and ARM64. An invalid choice will revert to the default.
IMAGE_SIZE This sets te '-s' flag which is used to make a custom sized image. The units are MB and the default is 8000. 8000 MB is the minimum size recommended for ALARM images, so the smallest that can be selected.

The /data/alarm-image directory is the working directory, and is where the image will be saved.

A sensible docker run command for Rock 64 Board might look like:

docker run -e BOARD=ROCK64 -v "/home/user/alarm-imager":/data/alarm-imager --privileged bowseruk/alarm-imager:latest

And for a armv7 build on a raspberry pi 3 would look like:

docker run -e BOARD=RASPBERRY_PI_3 -e ARCH=ARM -v "/home/user/alarm-imager":/data/alarm-imager --privileged bowseruk/alarm-imager:latest

The image will be saved in the Images directory of the working directory in the format ALARM-BOARD-ARCH e.g. ALARM-BANANA_PI-ARM.img. Most of the files used will be saved in the working directory to save time if a new image is made. You can delete these if this is not wanted.

To Do

There is some functionality I want to add to the script. This includes:

  • Adding a startup script with the recommended startup commands for each board.
  • Include a working script to resize boards to the full SD Card on first boot.
  • An interactive version of the script

Contribution

Feel free to send pull requests with any improvements. I only do this as a hobby, so changes may not be quick.