Welcome to Adam! Adam is the reference implementation of an LF-Edge API-compliant Controller. You can use Adam to drive one or more EVE instances from any computer, locally, in the cloud, or in a container.

Adam is a reference implementation. Thus, while it has all of the TLS encryption and authentication requirements of the official API, it has not been built or tested to withstand penetration attacks, DDoS, or any other major security requirements, and it has not been built with massive scale in mind. For those, please refer to various vendor cloud controller offerings, such as Zededa zedcloud.

To run Adam, you need a built Adam binary. Adam distributes in two modes:

• a single binary available on all major platforms: Linux, macOS, Windows. These are not yet available short of building it yourself (see below), but will be on the releases page when ready, at https://github.com/lf-edge/adam/releases/
• an OCI compliant container image available on Docker hub at https://hub.docker.com/r/lfedge/adam

The default command for the container is adam, so all adam commands can be run as either of the following:

$ adam <cmd>
$ docker run <cmd>

Every commit to the master branch of Adam releases an image to docker hub. These are tagged with the git commit. In addition, the most recent commit is tagged with snapshot, i.e. lfedge/adam:snapshot.

Adam regularly releases proper semantic versioned releases, with images tagged as vx.y.z, e.g. v1.2.3. The highest numbered, most recent release is tagged with latest.

• lfedge/adam:latest (or just lfedge/adam) - most recent release
• lfedge/adam:v1.2.3 - release 1.2.3
• lfedge/adam:snapshot - most recent commit to mainline branch
• lfedge/adam:6aa76a1ac3ee46aefd96525190e4bd4eb4f5d828 - build from commit 6aa76a1ac3ee46aefd96525190e4bd4eb4f5d828

For all of the sample commands in this guide, we use simply lfedge/adam, equivalent to lfedge/adam:latest. You should feel free to replace latest with whichever tag is appropriate for your use case.

In addition to adam itself - as a local binary or docker container - you need the following:

• A database directory, for the configs, data, and certificates. By default, this is under ./run/adam/ from the current directory, but is configurable; see the options.
• Server key and certificate for adam itself. By default this is ./run/adam/server-key.pem and ./run/adam/server.pem, but is configurable; see the options.

You can get the certificate and key before running adam as a server in one of several ways:

• pre-generate them yourself, using your favourite command: openssl, cfssl, etc.
• run adam generate
• run adam server --auto-cert, which will start up and generate the key and certificate, if it does not exist. You can specify CommonName of the certificate by setting the value of --cert-cn flag and hostnames and/or IP addresses by setting the value of --cert-hosts flag.
• provide the certificate in the SERVER_CERT environment variable and the key in the SERVER_KEY environment variable

The adam command has multiple options. The primary one is:

adam server

which will run Adam, listening on the default port of 8080 (it will tell you which when it starts), using the default server TLS key and certificate, using ./run/adam/ as its file storage location. These options are modifiable via the command-line; run adam server --help for options.

If you prefer to run Adam as a docker container:

docker run lfedge/adam server

You can add any of the options that would exist with a local Adam installation, including help: docker run lfedge/adam server --help.

Note that when running in a docker container, directories are ephemeral. If you want to keep the directories, you should bind-mount them into your container.

docker run -v $PWD/run:/somedir/run lfedge/adam server --conf-dir /somedir/run/config --db-url /somedir/run/adam --server-cert /somedir/run/adam/server.pem --server-key /somedir/run/server-key.pem

The default working directory for adam in the container is /adam/, which means the following will just work:

docker run -v $PWD/run:/adam/run lfedge/adam server

Or, you can use volume containers.

To make things easier, this repository includes a sample docker-compose.yml which runs adam, maps port 8080 in the container to 8080 on your host, and mounts the current directory's ./run/adam/ to the default /adam/run/adam/ in the container.

Finally, remember to map your ports when using a docker container:

docker run -v $PWD/run:/adam/run -p 8080:8080 lfedge/adam server

By default, adam listens on port 8080, but can be configured. Run adam server --help.

Finally, you can embed Adam container into an EVE root filesystem creating an EVE instance that can be controlled externally by clients talking to Adam and Adam relaying it to EVE. This comes very handy in testing and any other situation where turning EVE's configuration pull model into a push one makes sense. Note that this deployment mode forever commits a single Adam instance to a single EVE instance and all the communication between EVE and Adam happen via localhost on the running EVE edge node. Adam container has a script eve-embedded.sh that orchestrates this bond and the Adam container can be used to build EVE image via the following stanza in EVE's image YAML file:

   - name: adam
     image: lfedge/adam:latest
     binds:
        - /var/persist:/persist
        - /var/config:/config
     command: ["/bin/eve-embedded.sh"]
     net: host

Adam provides a CLI management interface and Web UI, both of which wrap an open management API.

Adam can be run in CLI mode, connecting to a local or remote Adam instance. The CLI does not manipulate the backend in any hidden way. It just sends REST commands over the well-defined management API.

To run Adam's CLI:

adam admin

Follow the options from there.

Adam currently features a basic Web UI, which is under active development. To access the Web UI, go to the main page for Adam, e.g. https://localhost:8080, which is the default.

Adam's Web server embeds all of its static files inside the binary, by default. There is an option to run Adam serving its files from the local filesystem. To do so, run:

adam server --web-dir <path-to-files>

For more information on developing the Web UI, see this document.

The management API is available at /admin. It currently is undocumented other than in the source code, but swagger is under development for it. Follow this issue.

Building Adam is straightforward:

1. Clone this repo
2. Ensure you have installed either go >= 1.16, or docker
3. Build

There are several options for building:

• make image will run the entire build in a docker container and give you a docker image. No local binaries will be created.
• make build will create the binary bin/adam-<os>-<arch> for your OS and architecture. You can override either by make OS=<os> and/or make ARCH=<arch>, e.g. make OS=linux ARCH=arm64. The build itself will happen with in a docker image or using your locally installed go; see below.
• make image-local will take the locally built binary and create an image with it.

These options allow you to do a one-step image build with no dependencies (make image), build a binary for your local usage (make build) using either your locally installed go or in a docker container, and make an image that uses your local binary (make image-local). The latter is often for quick reproducible builds.

All Makefile commands that execute go have the option to run locally or in docker. By default, they run in docker. If you prefer to run using a locally installed go, pass BUILD=local to any command, e.g.

make vet BUILD=local
make build BUILD=local

Adam requires TLS to communicate with EVE devices, which means a server key and certificate. If one is not available, it will fail startup. You can generate one using:

adam generate server

Run adam generate server --help for options. By default, it stores the server key and certificate in the same location as the default when running adam server.

For an EVE device to be accepted into Adam, it needs to be listed as one of:

• acceptable to onboard
• registered

An EVE device has to know the following before it can communicate with any controller (including Adam):

• controller's host name and port #
• controller's root certificate

additionally you may need to supply an entry mapping controller's host name to a routable IP address (in the /etc/hosts format)

When Adam server runs, it outputs all the required configuration in a folder specified by the conf-dir option (run/adam/config by default)

Onboarding is the process of enabling a device to self-register. This requires two pieces: an onboarding certificate, and a unique serial string. Each self-registering device must have a unique combination of onboarding certificate and serial string.

Adam has an onboarding directory where it maintains acceptable onboarding certificates and serials. By default, these are under ./run/adam/onboard/<cn>/, where the name cn is a file-friendly conversion of the certificate's Common Name. This directory contains two files:

• cert.pem - the actual onboarding certificate.
• serials.txt - a list of acceptable serials to use with this certificate, one per line. The wildcard * means any serial will be accepted.

You can modify these files directly; it is not, however, recommended.

Instead, use Adam's command-line admin options to work with the files:

adam generate

will generate a key/certificate pair, with a Common Name that you provide. Run adam generate --help for options.

You then can interact with the Adam server using adam admin:

$ adam admin device
$ adam admin onboard

These will list, add, remove, get or clear onboarding certificates and their serials, as well as devices directly.

Once you have generated an onboarding certificate, copy the certificate and key to the device to onboard.

More documentation is available in the docs/ directory.