Hardware Information API (hwapi)

The repo contains the API server and client for retrieving hardware information.

Prerequisites for deploying locally

• Install docker and setup permissions for server deployment
• Install rust and cargo for client deployment (https://www.rust-lang.org/tools/install)

Also, to run the server locally, make sure that another application doesn't use the port 8080.

Running API development server instance with Docker

Go to the server/ directory in the project. Since the server needs the data from the DB to work with, there are several options regarding how you can stand up the environment.

Using pre-populated database

You can use this option if you already have the SQLite DB file and want to use it. Make sure that the DB is located under the ./server/ directory.

You can stand up the environment by running the following commands:

# Assuming that the path to the DB file is ./hwapi.db
docker compose build --build-arg IMPORT_TOOL_PATH="" --build-arg DB_URL=sqlite:///./hwapi.db hwapi-dev
docker compose up --attach-dependencies hwapi-dev

Seed the database from the script

This approach doesn't require internet access and populates your DB with dummy data using the scripts/seed_db.py script.

export IMPORT_TOOL_PATH=./scripts/seed_db.py
docker compose up --attach-dependencies --build hwapi-dev

Load the data from C3

This approach populates the DB with the data from C3 (staging instance by default). Keep in mind that importing data from staging or production takes some time, you probably want to consider importing data from your local C3 instance with sample data.

To build and run the container with staging data, execute the following command:

docker compose up --attach-dependencies --build hwapi-dev

Alternatively, you can specify another C3 host (like production or the local one) by specifying the C3_URL:

export C3_URL=http://your.c3.instance  # e.g https://certification.canonical.com
docker compose up --attach-dependencies --build hwapi-dev

Accessing API schema

You can retrieve API schema in HTML, YAML, and JSON formats:

• The current version of the OpenAPI schema is browseable at canonical.github.io/hardware-api.
• To access the HTML view for the API schema, just run the server and follow the /#docs endpoint.
• Retrieve the schema in YAML from the running service by following the /openapi.yaml endpoint
• A copy of the openapi.yaml is included in the repo, and it is enforced by a CI automation to be up to date.
• For getting its JSON version, follow the /openapi.json endpoint.

Build the library (hwlib)

Currently, the library contains the functionality to collect the information about the machine hardware and OS, send the request to the server, and get the machine certification status.

sudo apt-get install -y pkgconf libssl-dev  # Install the required debian dependencies
cd client/hwlib
cargo build

Build and run the reference CLI tool (hwctl)

To check the machine certification status, run the following command. It sends the request to the server URL defined by HW_API_URL environment value (https://hw.ubuntu.com by default). The library requires root access since we collect the hardware information using SMBIOS data. If you're running it on a device that doesn't have SMBIOS data available, root privileges are not required.

cd client/
cargo build
sudo ./target/debug/hwctl

To send the request to a different server, run the tool specifying HW_API_URL environment variable:

cargo build
sudo HW_API_URL=https://your.server.url ./target/debug/hwctl

Building hwctl snap

To build and install hwctl as a snap locally, do the following after installing snapcraft and a build provider for it:

snapcraft --bind-ssh  # --verbose
sudo snap install ./hwctl_[version].snap --dangerous