Skip to content

Latest commit

 

History

History
225 lines (134 loc) · 10.1 KB

CONTRIBUTING.md

File metadata and controls

225 lines (134 loc) · 10.1 KB
Raw

ansible-collection

Required skills

In order to contribute to this Ansible collection, you will may need skills in:

  • Ansible
  • Python
  • Hyperledger Fabric
  • IBM Blockchain Platform
  • Docker
  • Red Hat OpenShift/Kubernetes
  • ReStructed Text
  • Jinja2 templates

The skills required depend on the area of the code you are looking to contribute, for example the Ansible modules are written in Python, the Ansible roles are written in YAML, and the documentation is written in ReStructed Text.

Setting up a development environment

Follow these steps to set up a development environment for this Ansible collection.

Note: that the repo is migrating touse Poetry for dependency management, and the just task runner for developmnet purposes

  1. Install Python v3.x. The Python version manager pyenv works great on Linux and macOS and avoids a lot of difficulties seen when trying to use the system Python, or Python installed using the package manager for your operating system.

We're migrating to use Poetry Optional for quickly running development activites use just

The linting uses shellcheck that needs to be installed with your package manager

  1. Clone the GitHub repository. To avoid difficulties later on, it is recommended that you clone the GitHub repository directly into the Ansible collections directory, so you need to delete any installed version first:

    rm -rf ~/.ansible/collections/ansible_collections/ibm/blockchain_platform

    Then clone the GitHub repository:

    git clone git@github.com:IBM-Blockchain/ansible-collection.git ~/.ansible/collections/ansible_collections/ibm/blockchain_platform

    You can then find the cloned GitHub repository at:

    ~/.ansible/collections/ansible_collections/ibm/blockchain_platform

    To make it easier to find, you may want to symlink it somewhere else:

    ln -s ~/.ansible/collections/ansible_collections/ibm/blockchain_platform ~/git/ansible-collection

  2. Open a terminal in the cloned GitHub repository, or change into the directory:

    cd ~/.ansible/collections/ansible_collections/ibm/blockchain_platform

  3. Install all Python dependencies:

    poetry install or pip install -Ur requirements.txt

    The modules installed include Ansible and tools for building documentation and linting the code.

  4. Install the Hyperledger Fabric tools. You can find the latest versions of these tools on the Hyperledger Fabric releases page - find the latest v2.x release, look for Assets, and download the appropriate file for your operating system.

    Once you have downloaded the appropriate file, it is recommended that you extract it into /usr/local so you don't have to set the PATH environment variable:

    tar xf hyperledger-fabric-darwin-amd64-2.3.1.tar.gz -C /usr/local

  5. In order to test any code changes you make to this collection, other than documentation only changes, you will need an IBM Blockchain Platform instance to test against.

    Depending on the change, you may need an instance of IBM Blockchain Platform running on IBM Cloud (SaaS), or the IBM Blockchain Platform software running in a Red Hat OpenShift or Kubernetes cluster, or both.

What's in the repository?

Ansible modules

Ansible modules, such as ibm.blockchain_platform.peer, are provided as Python modules under plugins/modules.

Each Ansible module has a main() function that is called by Ansible. At the start of each main() function, the arguments for the module are defined in a argument_spec, and then passed into Ansible by creating a BlockchainModule object, which is an instance of AnsibleModule.

Each Ansible module then goes through roughly the same process:

  1. Perform any additional input validation
  2. Connect to and determine the current state of the system
  3. Make any changes required to get the system into the defined state
  4. Return any data that may be of interest to the user

Ansible modules must return at the minimum a failure message using module.fail_json(), or a changed/not changed flag module.exit_json(changed=False).

Finally, each Ansible module has documentation strings such as ANSIBLE_METADATA, DOCUMENTATION, etc. These are parsed by the ansible-doc-extractor tool to generate the documentation for each module.

Python utility modules

Code that is used across multiple Ansible modules are provided as Python modules under plugins/module_utils.

Ansible roles

Ansible roles, such as ibm.blockchain_platform.endorsing_organization, can be found under roles.

Each Ansible role has a set of default parameters in defaults/main.yml, metadata in meta/main.yml, and the main set of tasks in tasks/main.yml. Most of the Ansible roles included in this collection have additional task files under the tasks directory that are called from tasks/main.yml.

Finally, there is no mechanism for generating documentation from an Ansible role. The documentation for the roles is written as ReStructured Text files under docs/source/roles.

Docker image

There is a Dockerfile at the root of the repository that builds a Docker image containing everything needed to use this Ansible collection, including the collection itself.

Integration tests

There are a few integration tests under tests/integration/targets. Integration tests in Ansible closely resemble Ansible roles, where there is a main set of tasks (including assertions) in tasks/main.yml.

Tutorial

The playbooks and configuration used in the build, join, and deploy tutorials is stored under tutorial. The legacy Hyperledger Fabric v1.x tutorial is stored under tutorial/v1.x, but this is never referenced from the documentation.

Documentation

The documentation is stored under the docs/source directory. There is a Jinja2 template used by ansible-doc-extractor under docs/templates.

Building the repository

Linting

You can lint the Ansible collection by running the following commands:

flake8 .
ansible-lint
for ROLE in roles/*; do ansible-lint ${ROLE}; done
for PLAYBOOK in tutorial/??-*.yml; do ansible-lint ${PLAYBOOK}; done
for PLAYBOOK in tutorial/v1.x/??-*.yml; do ansible-lint ${PLAYBOOK}; done
shellcheck tutorial/*.sh
yamllint .

This can also be run

just lint

Ansible collection

You can build the Ansible collection into a package by running the following command:

ansible-galaxy collection build

Other users can then install this package by running:

ansible-galaxy collection install ibm-blockchain_platform-1.0.0.tar.gz

The package can also be imported into Ansible Galaxy manually using your web browser.

All the above tasks can be completed by running:

just local

Documentation

The documentation can be built by running the following commands:

cd docs
make clean
make all

The built documentation will be available from the docs/build directory, open the index.html file in a web browser.

Testing

Integration tests

In order to run the integration tests, you must edit the file tests/integration/integration_config.yml and provide values for your IBM Blockchain Platform instance.

The test_run_id and short_test_run_id parameters provide namespacing for the tests that allow multiple instances of the tests to run concurrently on the same IBM Blockchain Platform instance. This is mostly useful in the automated CI/CD pipeline, but should not be needed if you are running against an individually owned IBM Blockchain Platform instance.

You can then run the integration tests by running the following command:

ansible-test integration

Tutorial tests

Running all of the tutorials is the best way to test the collection. The tutorials exercise most of the Ansible modules and roles in the collection.

CI

The CI for the collection is managed using GitHub Actions.

There is a main workflow defined in .github/workflows/main.yml. This is responsible for building the collection, linting the code, running the tests, and publishing the collection to Ansible Galaxy and Docker Hub when a release is created. This runs on every pull request, every merge into the main branch, every release/tag, and is also scheduled to run over night, every night.

There is also a "purge" workflow defined in .github/workflows/purge.yml. This is scheduled to run over night, every night, and simply purges all components from the IBM Blockchain Platform instances used by the main workflow. This is unfortunately required as sometimes the cleanup processes in the test do not occur or do not work, and the leftover components can take up valuable CPU, memory and storage on the associated Red Hat OpenShift cluster.

There are a large number of Action secrets defined here: https://github.com/IBM-Blockchain/ansible-collection/settings/secrets/actions

These include:

  • Ansible Galaxy API keys
  • Docker Hub API keys
  • IBM Blockchain Platform API keys
  • IBM Cloud API keys
  • Kubernetes information

Publishing/releasing

There are three parts to the publishing process:

All of this is automated using GitHub Actions.

To publish a new release, use GitHub by going to the Releases page.

For the tag version and release title, use the version number in galaxy.yml prefixed with v, for example v1.2.3.

When you publish the release, the GitHub Actions workflow will be triggered and it should automatically publish. You may need to restart the workflow if the tests fail due to intermittent issues such as slow or dropped network connections.

The workflow will automatically bump the version number in galaxy.yml after the publishing has completed.

Currently, sstone1, lesleyannjordan, and mbwhite have access to the ibm publisher on Ansible Galaxy.