You need to create a Github Organization: https://github.com/account/organizations/new

So far you need nothing but the organization name. Further steps while require additional data.

You need a master user, which is recommended not to be you main Github user, as you will need to provide an API Key with advanced permissions.

Create a new user on Github and then create an API Key with at least push, clone, including private repos, rights from: https://github.com/settings/tokens

Take not from that key, it will only show once.

Now, add this user to the organization you just created. That can be done from: https://github.com/orgs/YOUR-ORG/people, where "YOUR_ORG" is the name of the organization you created before.

Create the user-password combinations for the rabbit queues

printf <user> | docker secret create rabbit-user
printf <pass> | docker secret create rabbit-pass

• Python 3.x
• Docker (for mongo)

Install required Python libraries

pip install -r servers/webhooks/requirements.txt
pip install -r servers/frontend/requirements.txt

Create docker networks for test execution

docker network create --driver bridge backend
docker network create --internal results

Fire up mongo on Docker (note change /tmp/data if you want permanent storage)

docker run --name mongo -p 27017:27017 -v /tmp/data:/data/db --rm --network backend -d mongo

Build now (and everytime you change agents), build the dockers (you might need sudo)

./build.sh

Start the webhooks/tests server with

python start_service.py --github-api-key=<YOUR API KEY> --github-org=<YOUR ORGANIZATION NAME> --mongo-host=localhost

For example:

python start_service.py --port 8080 --github-fake-id=123 --no-github-init --github-api-key=231af25d0 --github-org=TNUI-UB --mongo-host=localhost

Processing tests synchronously instead of in a processes pool can be achieved by setting the environment variable DISABLE_POOL to 1, for example:

DISABLE_POOL=1 python start_service.py --github-api-key=231af25d0 --github-org=TNUI-UB --mongo-host=localhost

This proves to be useful when debugging Jobs or blocking for tests completion.

• Docker
• Docker-compose

First (and everytime you change something, ie. code), build the dockers (you might need sudo)

./build.sh

First time only, copy the file docker-compose.yml.template to docker-compose.yml and make sure to replace all <XXXX> values inside for its corresponding values.

Once it's ready, run it:

./run.sh

Each organization must have its own docker volume for the data. Suppose your organization name is "TNUI-UB" and your tests to be executed are "tests-assignment-2018". The following is to be done:

docker run -ti --rm --mount source=TNUI-UB-tests-assignment-2018,target=/tests alpine/git clone <TESTS_GIT_URL> /tests/_

Notice how the source argument is composed of <ORGANIZATION>-<TESTS>.