- Introduced new Immutable kind in the KFD API with associated functionalities.
- Implemented ClusterCreator, ClusterDeleter, CertificatesRenewer, and KubeconfigGetter for the Immutable kind.
- Added ExtraSchemaValidator for Immutable kind to validate schemas.
- Updated schema.go to include the new Immutable kind in the ExtraSchemaValidatorFactory.

…unction

- Introduced a new validation framework in `schema.go` to ensure the integrity of the Immutable KFD configuration.
- Added various validation checks including cluster name length, node configuration, uniqueness of hostnames, MAC addresses, and IP addresses.
- Implemented cross-reference checks for control plane, etcd, and node group members to ensure they reference existing nodes.
- Added network validation to check for CIDR overlaps and IP consistency with infrastructure nodes.
- Enhanced HA validation to provide warnings for sub-optimal configurations.
- Introduced a logging mechanism for validation warnings using logrus.
- Added utility functions for IP extraction and CIDR overlap checks in `ip.go`.

…figs

- Add `infrastructure_generate.go` to handle the generation of node configurations, including install and bootstrap Butane files, and their conversion to Ignition.
- Introduce templates for control plane, load balancer, and worker nodes in Butane format.
- Create a new `Infrastructure` struct in the `create` package to manage the infrastructure phase.
- Implement tests for the infrastructure generation logic, covering folder structure creation, template rendering, and config extraction.
- Embed Butane templates for easier access and rendering during the configuration generation process.

…ry directory

- Introduced new configuration management for infrastructure, merging user-defined settings with defaults.
- Added support for dynamic retrieval of SSH user, Flatcar version, and install disk settings from configuration.
- Enhanced the Prepare method to initialize configuration values and render Butane templates for nodes.
- Implemented a new method to split multi-document YAML files generated by Butane into individual files per node.
- Removed deprecated Butane templates for control plane, load balancer, and worker nodes, consolidating template management.
- Added a ToolValidator to ensure required tools are available for different operation phases.
- Updated the ClusterCreator to handle the new configuration structure and improved error handling.
- Created a sample configuration file (furyctl.yaml) for user reference.
- Added support for schema paths and validation for supported phases in the Immutable kind.

…ne and Ignition

…lates for Kubernetes

add a basic implementation of the serve command that starts an HTTP
server to serve assets from a defined path and waits for user
confirmation to stop

…tchbox rendering

…cture in tests

… publicKeyPath

- Implement tests for reading SSH public keys from specified paths.
- Cover scenarios including privateKeyPath, keyPath (deprecated), both specified, explicit publicKeyPath, derived publicKeyPath, and error cases.
- Ensure proper handling of empty public key files and non-existent public key files.
- Include environment variable expansion in path testing.

… and add tests

…tal notation

WIP: A base cabling of kubernetes, distribution and plugins phases for Immutable kind based on what kind OnPremises does.

Add missing pieces so the infrastructure phase for Immutable is run when running furyctl apply.

Add dummy file to the ugprades paths, otherwise furyctl fails, can be removed once we have an upgrade path

Add support for Immutable kind in the get supported-versions command.

Render table dynamically so we don't have to adjust manually the spacing when the data changes.

Update unit tests for the get supported-versions output to include Immutable kind.

Stop using the extractos for OnPrem and use the extractors for Immutable instead.

Add extractors for the Infrastructure phase.

Render files at the root of the infrastructure templates folder. Allows for example to render the boot.ipxe file.

Use better names for igntion files:
- bootstrap is now install-flatcar.json
- install is now node-config.json

Update directories, variables and function names accordingly

- Refactor infrastructure phase logic with new nomenclature for the butane (and ignition) files.
- Remove unused code.
- Remove unneeded methods.
- Improve logging.
- Linting.

Automatically continue when all nodes have booted instead of waiting for user confirmation.

…nd templates

- Pass also the flatcar version to the templates so they can build the right URL to download the Flatcar image.
- Make the Downloader client NOT decompress the downloaded files for the Flatcar image, otherwise the flatcar-install will try to deflate it again.
- Download the signature of the Flatcar image when downloading the image.

Remove unneeded helper wrapper funcs to check for binaries in cache when downloading assets. Our packages already do that.

Don't try to save the status to the cluster when running the infrastructure phase alone until the cluster has been created.

Ask for user confirmation when there are unsafe changes (unsafe reducers) to the infrastructure phase

Close waiting for user input channel when all nodes have booted otherwise we will try to close the server if the user presses enter even if it already has been closed when all nodes booted.

Write supported versions table to standard output instead of using logrus.Info.

Relates to: #651

Remove configData from the Infrastructure phase signature because we
don't need the merged configuration in this phase. The defaults that are
merged with the user data are only available for the distribution phase.

Reduce cyclomatic complexity in Creator logic of Immutable kind by
moving the confirmation logic to dedicated funcs

We want to have this code duplicated, because each provider can have
different schemas and support different phases, so having a common base
is a downside that could introduce changes in a provider that are not
wanted.

Add `**` as recursive matcher for the extractions rules.

This can be used to match changes on any of the child nodes of a path,
simplifying the number of rules needed.

Port some linting changes on error wrapping from
#659