arkdep attempts to be as simple to use as possible and avoid unnecessary abstraction, if you know how to use GNU/Linux picking up arkdep should be painless for it maintains much of your old familiar workflow.
- Systemd-boot bootloader is installed and configured as the primary bootloader
- dracut, wget and curl are installed
The following command will initialize arkdep, it will deploy a subvolume containing all arkdep related files excluding kernels and initramfs to `/arkdep`. Kernel and initramfs will instead be stored in `/boot/arkdep` upon generation.
Once ardep is installed you should prepare the overlay located at `/arkdep/overlay`. The overlay is copied directly on to the root filesystem of a new deployment, create directories inside of it as-if it were a root filesystem. For example, `/arkdep/overlay/etc` will be your `/etc` folder.
- passwd, shadow, group, subgid and subuid files containing only entries for root and normal user accounts, system accounts will be supplied via the images and are stored seperate in `/usr/lib`.
It will check in with the server defined in `/arkdep/config` as `repo_url` and pull the latest image defined in `$repo_url/$repo_default_image/database`, see [Repository](#Repository) for additional information.
arkdep will by default generate a psuedo-random hex string and use this as the name of your image. This behaviour can be overwritten by assigning a custom name to the `ARKDEP_CUSTOM_NAME` environment variable.
The `list` file is in part optional, it not utilized during the deployment process but the user may use it in combination with the `arkdep list` command to request a list of all available images in the repository.
The `database` file contains a `:` delimited list of all available images. Each line contains the following information `image_name:compression_method:sha1sum`.
arkdep will assume the top most entry in the database is the latest one, when no image version is defined or `latest` is requested it will grab the top most entry.