Labgrid Docker images

This folder contains Dockerfile's for building Docker images for the 3 different components of a Labgrid distributed infrastructure.

• labgrid-coordinator An image with the Labgrid coordinator. a Labgrid coordinator instance.
• labgrid-client An image with the Labgrid client tools and pytest integration.
• labgrid-exporter An image with the Labgrid exporter tools.

Build

To build one of the above images, you need to run the docker build command in the root of this repository. Example showing how to build labgrid-client image:

$ docker build --target labgrid-client -t docker.io/labgrid/client -f dockerfiles/Dockerfile .

Using BuildKit is recommended to reduce build times.

You can also choose to build all 3 images with the included script. The script will automatically use docker buildx <https://docs.docker.com/engine/reference/commandline/buildx/>` if available.

$ pip install --upgrade setuptools_scm
$ ./dockerfiles/build.sh

The script supports podman as well.

$ export DOCKER=podman
$ ./dockerfiles/build.sh

It builds for the native platform by default. However, building for foreign platforms is also supported by passing the platform(s) of choice, e.g. linux/arm64 as an additional argument.

$ pip install --upgrade setuptools_scm
$ ./dockerfiles/build.sh --platform linux/arm64

Usage

All 3 images are to be considered base images with the required software installed. No policy or configuration is done.

labgrid-coordinator usage

The labgrid-coordinator image can be used to run a coordinator instance.

It listens on port 20408, so you probably want to publish that so you can talk to the coordinator.

State is written to /opt/coordinator. You might want to bind a volume to that so you can restart the service without losing state.

$ docker run -t -p 20408:20408 -v $HOME/coordinator:/opt/coordinator \
      docker.io/labgrid/coordinator

labgrid-client usage

The labgrid-client image can be used to run labgrid-client and pytest commands. For example listing available places registered at coordinator at 192.168.1.42:20408

$ docker run -e LG_COORDINATOR=192.168.1.42:20408 docker.io/labgrid/client \
      labgrid-client places

Or running all pytest/labgrid tests at current directory:

$ docker run -e LG_COORDINATOR=192.168.1.42:20408 docker.io/labgrid/client \
      pytest

labgrid-exporter usage

The labgrid-exporter image runs a labgrid-exporter and optionally an ser2net service.

Configuration is not included, but needs to be bind mounted to /opt/conf/exporter.yaml and /opt/conf/ser2net.conf (optional).

Start it with something like:

$ docker run -e LG_COORDINATOR=192.168.1.42:20408 \
    -v $HOME/exporter-conf:/opt/conf \
      docker.io/labgrid/exporter

If using ser2net or if "exporting" e.g. a serial device, the devices needed must be added to Docker container (docker run --device option). Moreover, if using udev this must be mounted in as well: docker run -v run/udev:/run/udev:ro.

Staging

The staging folder contains a docker compose based example setup, where the images described above are used to create a setup with the following instances

• coordinator
• exporter
• client
• dut

The environment serves both to allow checking if the environment still function after changes, and can act as an example how to configure the docker images needed to run a minimal setup.

To use the staging environment to conduct a smoke test, first run docker compose to start all services except the client:

$ cd dockerfiles/staging
$ CURRENT_UID=$(id -u):$(id -g) docker compose up -d coordinator exporter dut

To run the smoke test just run the client:

$ docker compose up client