5.9 KiB
Build Container Installer Action
This action is used to enerate an ISO for installing an OSTree stored in a container image. This utilizes the anaconda command ostreecontainer
Usage
This action is designed to be called from a GitHub workflow using the following format
- name: Build ISO
uses: jasonn3/build-container-installer/v1.0.0
id: build
with:
arch: ${{ env.ARCH}}
image_name: ${{ env.IMAGE_NAME}}
image_repo: ${{ env.IMAGE_REPO}}
version: ${{ env.VERSION }}
variant: ${{ env.VARIANT }}
# This example is for uploading your ISO as a Github artifact. You can do something similar using any cloud storage, so long as you copy the output
- name: Upload ISO as artifact
id: upload
uses: actions/upload-artifact@v4
with:
name: my_iso.iso
path: |
${{ steps.build.outputs.iso_path }}
${{ steps.build.outputs.checksum-path }}
if-no-files-found: error
retention-days: 0
compression-level: 0
See Customizing for information about customizing the ISO that gets created using with
Customizing
The following variables can be used to customize the created ISO.
Inputs
| Variable | Description | Default Value |
|---|---|---|
| ARCH | Architecture for image to build | x86_64 |
| VERSION | Fedora version of installer to build | 39 |
| IMAGE_REPO | Repository containing the source container image | quay.io/fedora-ostree-desktops |
| IMAGE_NAME | Name of the source container image | base |
| IMAGE_TAG | Tag of the source container image | VERSION |
| EXTRA_BOOT_PARAMS | Extra params used by grub to boot the anaconda installer | [empty] |
| VARIANT | Source container variant* | Server |
| WEB_UI | Enable Anaconda WebUI (experimental) | false |
| ISO_NAME | Name of the ISO you wish to output when completed | IMAGE_NAME-IMAGE_TAG |
Available options for VARIANT can be found by running dnf provides system-release.
Variant will be the third item in the package name. Example: fedora-release-kinoite-39-34.noarch will be kinoite
Outputs
| Variable | Description | Usage |
|---|---|---|
| iso-path | Path to ISO file that the action creates | ${{ steps.YOUR_ID_FOR_ACTION.outputs.iso_path }} |
| checksum-path | Path to the checksum file that the action creates | ${{ steps.YOUR_ID_FOR_ACTION.outputs.iso_path }} |
For outputs, see example above.
Development
Makefile
The Makefile contains all of the commands that are run in the action. There are separate targets for each file generated, however make can be used to generate the final image and make clean can be used to clean up the workspace. The resulting ISO will be stored in the build directory.
make install-deps can be used to install the necessary packages
See Customizing for information about customizing the ISO that gets created.
Container
A container with make install-deps already run is provided at ghcr.io/jasonn3/build-container-installer:latest
To use the container file, run docker run --privileged --volume .:/build-container-installer/build ghcr.io/jasonn3/build-container-installer:latest.
This will create an ISO with the baked in defaults of the container image.
See Customizing for information about customizing the ISO that gets created. The variable can either be defined as environment variables. Examples:
Building an ISO to install Fedora 38
docker run --rm --privileged --volume .:/build-container-installer/build ghcr.io/jasonn3/build-container-installer:latest VERSION=38 IMAGE_NAME=base IMAGE_TAG=38 VARIANT=Server
Building an ISO to install Fedora 39
docker run --rm --privileged --volume .:/build-container-installer/build ghcr.io/jasonn3/build-container-installer:latest VERSION=39 IMAGE_NAME=base IMAGE_TAG=39 VARIANT=Server
VSCode Dev Container
There is a dev container configuration provided for development. By default it will use the existing container image available at ghcr.io/jasonn3/build-container-installer:latest, however, you can have it build a new image by editing .devcontainer/devcontainer.json and replacing image with build. Ctrl+/ can be used to comment and uncomment blocks of code within VSCode.
The code from VSCode will be available at /workspaces/build-container-installer once the container has started.
Privileged is required for access to loop devices for lorax.
Use existing container image:
{
"name": "Existing Dockerfile",
// "build": {
// "context": "..",
// "dockerfile": "../Containerfile",
// "args": {
// "version": "39"
// }
// },
"image": "ghcr.io/jasonn3/build-container-installer:latest",
"overrideCommand": true,
"shutdownAction": "stopContainer",
"privileged": true
}
Build a new container image:
{
"name": "Existing Dockerfile",
"build": {
"context": "..",
"dockerfile": "../Containerfile",
"args": {
"version": "39"
}
},
//"image": "ghcr.io/jasonn3/build-container-installer:latest",
"overrideCommand": true,
"shutdownAction": "stopContainer",
"privileged": true
}