Skip to content

Latest commit

 

History

History
409 lines (313 loc) · 15.8 KB

DEVELOPER.md

File metadata and controls

409 lines (313 loc) · 15.8 KB
Raw

Development

We encourage community contributions to Kong. To make sure it is a smooth experience (both for you and for the Kong team), please read CONTRIBUTING.md, CODE_OF_CONDUCT.md, and COPYRIGHT before you start.

If you are planning on developing on Kong, you'll need a development installation. The master branch holds the latest unreleased source code.

You can read more about writing your own plugins in the Plugin Development Guide, or browse an online version of Kong's source code documentation in the Plugin Development Kit (PDK) Reference.

For a quick start with custom plugin development, check out Pongo and the plugin template explained in detail below.

Distributions

Kong comes in many shapes. While this repository contains its core's source code, other repos are also under active development:

You can find every supported distribution on the official installation page.

Kong Pongo

Pongo is a CLI tool that are specific for plugin development. It is docker-compose based and will create local test environments including all dependencies. Core features are running tests, integrated linter, config initialization, CI support, and custom dependencies.

Kong Plugin Template

The plugin template provides a basic plugin and is considered a best-practices plugin repository. When writing custom plugins, we strongly suggest you start by using this repository as a starting point. It contains the proper file structures, configuration files, and CI setup to get up and running quickly. This repository seamlessly integrates with Pongo.

Build and Install from source

This is the hard way to build a development environment, and also a good start for beginners to understand how everything fits together.

Kong is mostly an OpenResty application made of Lua source files, but also requires some additional third-party dependencies, some of which are compiled with tweaked options, and kong runs on a modified version of OpenResty with patches.

To install from the source, first, we clone the repository:

git clone https://github.com/Kong/kong

cd kong
# You might want to switch to the development branch. See CONTRIBUTING.md
git checkout master

Then we will install the dependencies:

Ubuntu/Debian:

sudo apt update \
&& sudo apt install -y \
    automake \
    build-essential \
    curl \
    cmake \
    file \
    git \
    libyaml-dev \
    libprotobuf-dev \
    m4 \
    perl \
    pkg-config \
    procps \
    unzip \
    valgrind \
    zlib1g-dev

Fedora/CentOS/RHEL:

dnf install \
    automake \
    cmake \
    gcc \
    gcc-c++ \
    git \
    libyaml-devel \
    make \
    patch \
    perl \
    protobuf-devel \
    unzip \
    valgrind \
    valgrind-devel \
    zlib-devel

macOS

# Install XCode instead of Command Line Tools is recommended
xcode-select --install
# Install HomeBrew
/bin/bash -c "$(curl -fsSL https://raw.githubusercontent.com/Homebrew/install/HEAD/install.sh)"
# Build dependencies
brew install libyaml cmake

cmake 3 is needed to build some targets, some distributions ship version 2 only. An updated cmake can be downloaded here.

Finally, we start the build process:

# Build the virutual environment for developing Kong
make build-venv

The build guide contains a troubleshooting section if you face any problems. It also describes the build process in detail, if you want to development on the build system itself.

Databases

The easiest way to handle these as a single group is via docker-compose. It's also recommended to set your user as a docker manager to simplify the next steps.

Make sure the docker daemon is enabled and running: sudo systemctl enable docker and sudo systemctl start docker. Verify that docker ps shows no errors.

On a Fedora VM, you might have to disable SELinux:

sudo vim /etc/selinux/config        # change the line to SELINUX=disabled
sudo setenforce 0

Now pull the compose script from the repository and fire it up:

git clone https://github.com/thibaultcha/kong-tests-compose.git
cd kong-tests-compose
docker-compose up

Verify the three new containers are up and running with docker ps on a separate terminal.

Start Kong

Now you can start Kong:

# active the venv into your shell envirnoment
# For Zsh/Bash:
. bazel-bin/build/kong-dev-venv.sh
# For Fish Shell:
. bazel-bin/build/kong-dev-venv.fish
# Start Kong!
kong start

Install Development Dependencies

Running for development

By default, the development environment adds current directory to Lua files search path.

Modifying the lua_package_path and lua_package_cpath directives will allow Kong to find your custom plugin's source code wherever it might be in your system.

Tests

Install the development dependencies ([busted], [luacheck]) with:

make setup-dev-env

Kong relies on three test suites using the [busted] testing library:

  • Unit tests
  • Integration tests, which require Postgres and Cassandra to be up and running
  • Plugins tests, which require Postgres to be running

The first can simply be run after installing busted and running:

make test

However, the integration and plugins tests will spawn a Kong instance and perform their tests against it. Because these test suites perform their tests against the Kong instance, you may need to edit the spec/kong_tests.conf configuration file to make your test instance point to your Postgres/Cassandra servers, depending on your needs.

You can run the integration tests (assuming both Postgres and Cassandra are running and configured according to spec/kong_tests.conf) with:

make test-integration

And the plugins tests with:

make test-plugins

Finally, all suites can be run at once by simply using:

make test-all

Consult the run_tests.sh script for more advanced example usage of the test suites and the Makefile.

Finally, a very useful tool in Lua development (as with many other dynamic languages) is performing static linting of your code. You can use [luacheck] (installed with make dev) for this:

make lint

Upgrade tests

Kong Gateway supports no-downtime upgrades through its database schema migration mechanism (see UPGRADE.md). Each schema migration needs to be written in a way that allows the previous and the current version of Kong Gateway run against the same database during upgrades. Once all nodes have been upgraded to the current version of Kong Gateway, additional changes to the database can be made that are incompatible with the previous version. To support that, each migration is split into two parts, an up part that can only make backwards-compatible changes, and a teardown part that runs after all nodes have been upgraded to the current version.

Each migration that is contained in Kong Gateway needs to be accompanied with a test that verifies the correct operation of both the previous and the current version during an upgrade. These tests are located in the spec/05-migration/ directory and must be named after the migration they test such that the migration kong/**/*.lua has a test in spec/05-migration/**/*_spec.lua. The presence of a test is enforced by the upgrade testing shell script which is automatically run through a GitHub Action.

The upgrade testing shell script works as follows:

  • A new Kong Gateway installation is brought up using Gojira, consisting of one node containing the previous version of Kong Gateway ("OLD"), one node containing the current version of Kong Gateway ("NEW") and a shared database server (PostgreSQL or Cassandra).
  • NEW: The database is initialized using kong migrations bootstrap.
  • OLD: The setup phase of all applicable migration tests is run.
  • NEW: kong migrations up is run to run the up part of all applicable migrations.
  • OLD: The old_after_up phase of all applicable migration tests is run.
  • NEW: The new_after_up phase of all applicable migration tests is run.
  • NEW: kong migrations finish is run to invoke the teardown part of all applicable migrations.
  • NEW: The new_after_finish phase of all applicable migration tests is run.

Upgrade tests are run using [busted]. To support the specific testing method of upgrade testing, a number of helper functions are defined in the spec/upgrade_helpers.lua module. Migration tests use functions from this module to define test cases and associate them with phases of the upgrade testing process. Consequently, they are named setup, old_after_up, new_after_up and new_after_finish. Additonally, the function all_phases can be used to run a certain test in the three phases old_after_up, new_after_up and new_after_finish. These functions replace the use of busted's it function and accept a descriptive string and a function as argument.

It is important to note that upgrade tests need to run on both the old and the new version of Kong. Thus, they can only use features that are available in both versions (i.e. from helpers.lua). The module spec/upgrade_helpers.lua is copied from the new version into the container of the old version and it can be used to make new library functionality available to migration tests.

Makefile

When developing, you can use the Makefile for doing the following operations:

Name Description
install Install the Kong luarock globally
dev Install development dependencies
lint Lint Lua files in kong/ and spec/
test Run the unit tests suite
test-integration Run the integration tests suite
test-plugins Run the plugins test suite
test-all Run all unit + integration + plugins tests at once

Dev on Linux (Host/VM)

If you have a Linux development environment (either virtual or bare metal), the build is done in four separate steps:

  1. Development dependencies and runtime libraries, including:
    1. Prerequisite packages. Mostly compilers, tools, and libraries required to compile everything else.
    2. OpenResty system, including Nginx, LuaJIT, PCRE, etc.
  2. Databases. Kong uses Postgres, Cassandra, and Redis. We have a handy setup with docker-compose to keep each on its container.
  3. Kong itself.

Virtual Machine (Optional)

Final deployments are typically on a Linux machine or container, so even if all components are multiplatform, it's easier to use it for development too. If you use macOS or Windows machines, setting up a virtual machine is easy enough now. Most of us use the freely available VirtualBox without any trouble.

If you use Linux for your desktop, you can skip this section.

There are no "hard" requirements on any Linux distro, but RHEL and CentOS can be more of a challenge to get recent versions of many packages; Fedora, Debian, or Ubuntu are easier for this.

To avoid long compilation times, give the VM plenty of RAM (8GB recommended) and all the CPU cores you can.

Virtual Box setup

You will need to setup port forwarding on VirtualBox to be able to ssh into the box which can be done as follows:

  1. Select the virtual machine you want to use and click "Settings"
  2. Click the "Network" tab
  3. Click the "Advanced" dropdown
  4. Click "Port Forwarding"
  5. Add a new rule in the popup. The only thing you will need is "Host Port" to be 22222 and "Guest Port" to be 22. Everything else can be left default (see screenshot below)
  6. Click "Ok"

Now you should be able to ssh <your_name>@127.1 -p 22222 to get SSH prompt. However, this requires us to type a long command and password every time we sign in. It is recommended you set up a public key and SSH alias to make this process simpler:

  1. On your host machine, generate a keypair for SSH into the guest: ssh-keygen -t ed25519. Just keep hitting Enter until the key is generated. You do not need a password for this key file since it is only used for SSH into your guest
  2. Type cat .ssh/id_ed25519.pub and copy the public key
  3. SSH into the guest using the command above
  4. Create the ssh config directory (if it doesn't exist) $ mkdir -p .ssh
  5. Edit the authorized keys list: vim .ssh/authorized_keys
  6. Paste in the content of .ssh/id_ed25519.pub
  7. Adjust the required privileges: chmod 700 .ssh/ and chmod 400 .ssh/authorized_keys
  8. Logout of guest and make sure you are not promoted password when SSH again
  9. Edit the .ssh/config file on your host and put in the following content:
    Host dev
        HostName 127.1
        Port 22222
        User <your_user_name>

Now try ssh dev on your host, you should be able to get into the guest directly.

Dev on VSCode Container / GitHub Codespaces

The devcontainer.json file in Kong's project tells VS Code how to access (or create) a development container with a well-defined tool and runtime stack.

What's next