Learning locker docker images

This repository contains the Dockerfile to build images for:

learning locker, a LRS to store and analyze xAPI statements.
xapi-service, a service dedicated to xAPI requests.

We follow learning locker's versions, so images built from this repository will have the same tags you can found in learning locker and xapi-service repositories.

Building learning locker image

To build a learning locker image, you have to pick the right tag to use and then build it using the following docker command (here with the v2.6.2 tag):

$ docker build -t fundocker/learninglocker:v2.6.2 --build-arg LL_VERSION="v2.6.2" learninglocker/

Building xapi-service image

To build the xapi-service image, you have to pick the right tag to use and then build it using the following docker command (here with the v2.2.15 tag):

$ docker build -t fundocker/xapi-service:v2.2.15 --build-arg VERSION="v2.2.15" xapi/

How to use the CI

The CI configured in this repository will test if the image can be built successfully for a particular release, but it also pushes images to docker hub when this repository is tagged

Here are the steps to follow to publish a new image:

Edit the .circleci/releases.sh file and update the tag version corresponding to the image you want to publish.
Commit your changes, submit a pull request and once merged into master you will be able to tag a new version.

Tags must stick to the following patterns for the image release you want to publish:

learninglocker-[RELEASE_TAG] if you want to publish a new learning locker image (e.g. learninglocker-v2.6.2),
xapi-service-[RELEASE_TAG] if you want to publish a new xapi-service image (e.g. xapi-service-v2.2.15).

Using the release script

If you have more than one release to publish, we also provide the bin/release Bash script to automate the whole releasing process.

Requirements

The bin/release script requires to install the two following dependencies on your system:

curl
jq

You will find a packaged version of those tools on most *nix systems.

Usage

First step: query the public GitHub API to check available releases of the LearningLocker and xapi-service projects that have not been committed yet and create a version upgrade commit per release in a new branch. This can be achieved thanks to the prepare command:

$ bin/release prepare

Check that missing releases have been committed in your current branch using the git log command.

If everything goes well, you can push you local branch to GitHub to open a new pull request:

$ bin/release push

Note that we will not push the local branch at once, but commit per commit to run the CI on each commit and ensure that the build passes for all of them.

Once your pull request has been opened, reviewed and merged to master, you must tag each commit with the appropriate release tag and push this tag to GitHub to trigger the final image build and publication to DockerHub:

$ bin/release tag

Now you just have to wait for the CI to do its job.

nota bene: the bin/release script accepts a -t (--token) option to perform authenticated requests against GitHub's public API. This allows you to bypass the 60 requests per hour rate limit. In this case, you will need to generate a new personnal access token from GitHub's interface and use it as follows:

$ bin/release -t THETOKEN CMD

Using the learning locker image

You have to configure your learning locker image using environment variables. You will find all available variables in the project documentation.

There is no entrypoint nor command configured in this image. You will have to declare them explicitly when you run your container. A list of commands to use per service follows:

Running the UI application:

$ docker run --rm fundocker/learninglocker:v2.6.2 node ui/dist/server

Running the API application:

$ docker run --rm fundocker/learninglocker:v2.6.2 node api/dist/server

Running the worker application:

$ docker run --rm fundocker/learninglocker:v2.6.2 node worker/dist/server

You can also use the CLI by executing the following command from your container:

$ docker run --rm fundocker/learninglocker:v2.6.2 node cli/dist/server

Using the xapi-service image

You have to configure your xapi-service image using environment variables. You will find all available variables in the project documentation.

By default, the entrypoint from this image will run the xapi-service, so you don't have to declare a command.

docker-compose example

In the example directory you will find a docker-compose configuration. This configuration is here as an example to help you to integrate Learning Locker in your stack.

You can run it without modification in this directory, you can test it by connecting to the interface with the url http://localhost:8080. You will also have to create a site admin by executing this command once your docker-compose is up and running:

$ docker-compose exec api node cli/dist/server createSiteAdmin [email] [organisation] [password]

You will probably want to integrate the Learning Locker API in your project, to do that you must copy the docker-compose.yml content in your docker-compose.yml project file and adapt it to match with your stack.

License

This work is released under the MIT License (see LICENSE).

Async0x42 / learninglocker-docker