A container to create disk images from bootc container inputs, especially oriented towards Fedora/CentOS bootc or derivatives.
Have podman installed on your system. Either through your systems package manager if you're on Linux or through Podman Desktop if you are on macOS or Windows. If you want to run the resulting virtual machine(s) or installer media you can use qemu.
A very nice GUI extension for Podman Desktop is also available. The command line examples below can be all handled by Podman Desktop.
On macOS, the podman machine must be running in rootful mode:
$ podman machine stop # if already running
Waiting for VM to exit...
Machine "podman-machine-default" stopped successfully
$ podman machine set --rootful
$ podman machine start
The following example builds a centos-bootc:stream9
bootable container into a QCOW2 image for the architecture you're running
the command on. However, be sure to see the upstream documentation
for more general information! Note that outside of initial experimentation, it's recommended to build a derived container image
(or reuse a derived image built via someone else) and then use this project to make a disk image from your custom image.
The generic base images do not include a default user. This example injects a user configuration file by adding a volume-mount for the local file to the bootc-image-builder container.
The following command will create a QCOW2 disk image. First, create ./config.toml
as described above to configure user access.
# Ensure the image is fetched
sudo podman pull quay.io/centos-bootc/centos-bootc:stream9
sudo podman run \
--rm \
-it \
--privileged \
--pull=newer \
--security-opt label=type:unconfined_t \
-v $(pwd)/config.toml:/config.toml \
-v $(pwd)/output:/output \
-v /var/lib/containers/storage:/var/lib/containers/storage \
quay.io/centos-bootc/bootc-image-builder:latest \
--type qcow2 \
--local \
quay.io/centos-bootc/centos-bootc:stream9
A virtual machine can be launched using qemu-system-x86_64
or with virt-install
as shown below;
however there is more information about virtualization and other
choices in the Fedora/CentOS bootc documentation.
qemu-system-x86_64 \
-M accel=kvm \
-cpu host \
-smp 2 \
-m 4096 \
-bios /usr/share/OVMF/OVMF_CODE.fd \
-serial stdio \
-snapshot output/qcow2/disk.qcow2
sudo virt-install \
--name fedora-bootc \
--cpu host \
--vcpus 4 \
--memory 4096 \
--import --disk ./output/qcow2/disk.qcow2,format=qcow2 \
--os-variant fedora-eln
This assumes qemu was installed through homebrew.
qemu-system-aarch64 \
-M accel=hvf \
-cpu host \
-smp 2 \
-m 4096 \
-bios /opt/homebrew/Cellar/qemu/8.1.3_2/share/qemu/edk2-aarch64-code.fd \
-serial stdio \
-machine virt \
-snapshot output/qcow2/disk.qcow2
Usage:
sudo podman run \
--rm \
-it \
--privileged \
--pull=newer \
--security-opt label=type:unconfined_t \
-v $(pwd)/output:/output \
quay.io/centos-bootc/bootc-image-builder:latest \
<imgref>
Flags:
--chown string chown the ouput directory to match the specified UID:GID
--tls-verify require HTTPS and verify certificates when contacting registries (default true)
--type string image type to build [qcow2, ami] (default "qcow2")
--target-arch string architecture to build image for (default is the native architecture)
Argument | Description | Default Value |
---|---|---|
--chown | chown the output directory to match the specified UID:GID | β |
--rootfs | Root filesystem type. Overrides the default from the source container. Supported values: ext4, xfs | |
--tls-verify | Require HTTPS and verify certificates when contacting registries | true |
--type | Image type to build | qcow2 |
--target-arch | Target arch to build | β |
The --type
parameter can be given multiple times and multiple outputs will
be produced.
π‘ Tip: Flags in bold are the most important ones.
The following image types are currently available via the --type
argument:
Image type | Target environment |
---|---|
ami |
Amazon Machine Image |
qcow2 (default) |
QEMU |
vmdk |
VMDK usable in vSphere, among others |
anaconda-iso |
An unattended Anaconda installer that installs to the first disk found. |
raw |
Unformatted raw disk. |
Specify the target architecture of the system on which the disk image will be installed on. By default,
bootc-image-builder
will build for the native host architecture. The target architecture
must match an available architecture of the bootc-image-builder
image you are using to build the disk image.
Currently, amd64
and arm64
are included in quay.io/centos-bootc/bootc-image-builder
manifest list.
The architecture of the bootc OCI image and the bootc-image-builder image must match. For example, when building
a non-native architecture bootc OCI image, say, building for x86_64 from an arm-based Mac, it is possible to run
podman build
with the --platform linux/amd64
flag. In this case, to then build a disk image from the same arm-based Mac,
you should provide --target-arch amd64
when running the bootc-image-builder
command.
In order to successfully import an AMI into your AWS account, you need to have the vmimport service role configured on your account.
AMIs can be automatically uploaded to AWS by specifying the following flags:
Argument | Description |
---|---|
--aws-ami-name | Name for the AMI in AWS |
--aws-bucket | Target S3 bucket name for intermediate storage when creating AMI |
--aws-region | Target region for AWS uploads |
Notes:
- These flags must all be specified together. If none are specified, the AMI is exported to the output directory.
- The bucket must already exist in the selected region, bootc-image-builder will not create it if it is missing.
- The output volume is not needed in this case. The image is uploaded to AWS and not exported.
If you already have a credentials file (usually in $HOME/.aws/credentials
) you need to forward the
directory to the container
For example:
$ sudo podman run \
--rm \
-it \
--privileged \
--pull=newer \
--security-opt label=type:unconfined_t \
-v $HOME/.aws:/root/.aws:ro \
--env AWS_PROFILE=default \
quay.io/centos-bootc/bootc-image-builder:latest \
--type ami \
--aws-ami-name centos-bootc-ami \
--aws-bucket fedora-bootc-bucket \
--aws-region us-east-1 \
quay.io/centos-bootc/centos-bootc:stream9
Notes:
- you can also inject ALL your AWS configuration parameters with
--env AWS_*
see the AWS CLI documentation for more information about other environment variables
AWS credentials can be specified through two environment variables:
Variable name | Description |
---|---|
AWS_ACCESS_KEY_ID | AWS access key associated with an IAM account. |
AWS_SECRET_ACCESS_KEY | Specifies the secret key associated with the access key. This is essentially the "password" for the access key. |
Those should not be specified with --env
as plain value, but you can silently hand them over with --env AWS_*
or
save these variables in a file and pass them using the --env-file
flag for podman run
.
For example:
$ cat aws.secrets
AWS_ACCESS_KEY_ID=AKIAIOSFODNN7EXAMPLE
AWS_SECRET_ACCESS_KEY=wJalrXUtnFEMI/K7MDENG/bPxRfiCYEXAMPLEKEY
$ sudo podman run \
--rm \
-it \
--privileged \
--pull=newer \
--security-opt label=type:unconfined_t \
--env-file=aws.secrets \
quay.io/centos-bootc/bootc-image-builder:latest \
--type ami \
--aws-ami-name centos-bootc-ami \
--aws-bucket centos-bootc-bucket \
--aws-region us-east-1 \
quay.io/centos-bootc/centos-bootc:stream9
The following volumes can be mounted inside the container:
Volume | Purpose | Required |
---|---|---|
/output |
Used for storing the resulting artifacts | β |
/store |
Used for the osbuild store | No |
/rpmmd |
Used for the DNF cache | No |
A build config is a Toml (or JSON) file with customizations for the resulting image. The config file is mapped into the container directory to /config.toml
. The customizations are specified under a customizations
object.
As an example, let's show how you can add a user to the image:
Firstly create a file ./config.toml
and put the following content into it:
[[customizations.user]]
name = "alice"
password = "bob"
key = "ssh-rsa AAA ... user@email.com"
groups = ["wheel"]
Then, run bootc-image-builder
with the following arguments:
sudo podman run \
--rm \
-it \
--privileged \
--pull=newer \
--security-opt label=type:unconfined_t \
-v $(pwd)/config.toml:/config.toml \
-v $(pwd)/output:/output \
quay.io/centos-bootc/bootc-image-builder:latest \
--type qcow2 \
quay.io/centos-bootc/centos-bootc:stream9
Possible fields:
Field | Use | Required |
---|---|---|
name |
Name of the user | β |
password |
Unencrypted password | No |
key |
Public SSH key contents | No |
groups |
An array of secondary to put the user into | No |
Example:
{
"user": [
{
"name": "alice",
"password": "bob",
"key": "ssh-rsa AAA ... user@email.com",
"groups": [
"wheel",
"admins"
]
}
]
}
{
"customizations": {
"kernel": {
"append": "mitigations=auto,nosmt"
}
}
}
To build the container locally you can run
sudo podman build --tag bootc-image-builder .
NOTE: running already the podman build
as root avoids problems later as we need to run the building
of the image as root anyway
With a virtual machine launched with the above virt-install example, access the system with
ssh -i /path/to/private/ssh-key alice@ip-address
Note that if you do not provide a password for the provided user, sudo
will not work unless passwordless sudo
is configured. The base image quay.io/centos-bootc/centos-bootc:stream9
does not configure passwordless sudo.
This can be configured in a derived bootc container by including the following in a Containerfile.
FROM quay.io/centos-bootc/centos-bootc:stream9
ADD wheel-passwordless-sudo /etc/sudoers.d/wheel-passwordless-sudo
The contents of the file $(pwd)/wheel-passwordless-sudo
should be
%wheel ALL=(ALL) NOPASSWD: ALL
- Website: https://www.osbuild.org
- Bug Tracker: https://github.com/osbuild/bootc-image-builder/issues
- Matrix: #image-builder on fedoraproject.org
- Mailing List: image-builder@redhat.com
- Changelog: https://github.com/osbuild/bootc-image-builder/releases
Please refer to the developer guide to learn about our workflow, code style and more.
- web: https://github.com/osbuild/bootc-image-builder
- https:
https://github.com/osbuild/bootc-image-builder.git
- ssh:
git@github.com:osbuild/bootc-image-builder.git
- Apache-2.0
- See LICENSE file for details.