Skip to content

Latest commit

 

History

History
177 lines (111 loc) · 7.07 KB

getting-started.md

File metadata and controls

177 lines (111 loc) · 7.07 KB

Getting started

This document describes how to setup your development environment.

Preparation

Make sure the following software is installed and added to the $PATH variable:

Clone the repository and install the dependencies:

yarn

Running the cluster

To make Dashboard work you need to have cluster running. If you would like to use local cluster we recommend kubeadm, minikube or kubeadm-dind-cluster. The most convenient way is to make it work is to create a proxy. Run the following command:

kubectl proxy

kubectl will handle authentication with Kubernetes and create an API proxy with the address localhost:8080. Therefore, no changes in the configuration are required.

Serving Dashboard for Development

Quick updated version:

npm start

Another way to connect to real cluster while developing dashboard is to specify options for npm like following:

npm run start:https --kubeconfig=<path to your kubeconfig>

Please see here which options you can specify to run dashboard with npm.

Open a browser and access the UI under localhost:8080.

In the background, npm start makes a concurrently call to start the golang backend server and the angular development server.

Once the angular server starts, it takes some time to pre-compile all assets before serving them. By default, the angular development server watches for file changes and will update accordingly.

As stated in the Angular documentation, i18n does not work in the development mode. Follow Building Dashboard for Production section to test this feature.

Due to the deployment complexities of i18n and the need to minimize rebuild time, the development server only supports localizing a single locale at a time. Setting the "localize" option to true will cause an error when using ng serve if more than one locale is defined. Setting the option to a specific locale, such as "localize": ["fr"], can work if you want to develop against a specific locale (such as fr).

Building Dashboard for Production

To build dashboard for production, you still need to install bc.

The Dashboard project can be built for production by using the following task:

make build

The code is compiled, compressed, i18n support is enabled and debug support removed. The dashboard binary can be found in the dist folder.

To build and immediately serve Dashboard from the dist folder, use the following task:

make prod

Open a browser and access the UI under localhost:9090. The following processes should be running (respective ports are given in parentheses):

Dashboard backend (9090) ---> Kubernetes API server (8080)

To build the docker image on darwin OS you will need to set environment variable for go to build as linux:

export GOOS=linux

In order to package everything into a ready-to-run Docker image, use the following task:

make docker-build-head

You might notice that the Docker image is very small and requires only a few MB. Only Dashboard assets are added to a scratch image. This is possible, because the dashboard binary has no external dependencies. Awesome!

Run the tests

Unit tests should be executed after every source code change. The following task makes this a breeze. The full test suite includes unit tests and integration tests.

make test

You can also run individual tests on their own (such as the backend or frontend tests) by doing the following:

make test-backend
make test-frontend

The code style check suite includes format checks can be executed with:

make check

The code formatting can be executed with:

make fix

These check and formatting involves in go, ts, scss, html, license and i18n files.

Committing changes to your fork

Before committing any changes, please run make fix. This will keep you from accidentally committing non tested and unformatted code.

Since the hooks for commit has been set with husky into <dashboard_home>/.git/hooks/pre-commit already if you installed dashboard according to above, so it will run make fix and keep your code as formatted.

Then you can commit your changes and push them to your fork:

git commit
git push -f origin my-feature

Easy way to build your development environment with Docker

At first, change directory to kubernetes dashboard repository of your fork. This development container has all of dependencies to develop dashboard.

Miscellaneous

  • Development container builds Kubernetes Dashboard and runs it with self-certificates by default.
  • This container create user with UID and GID same as local user, switch user to user with gosu and run commands. So created or updated files like results of make fix would have same ownership as your host. You can commit them immediately from your host.
  • Built Kubernetes Dashboard will run by docker compose, so other few containers will be created in your docker.

To run with your Kubernetes cluster

  1. Copy kubeconfig from your cluster, and confirm the URL for API server in it, and modify it if necessary.
  2. Set filepath for kubeconfig into K8S_DASHBOARD_KUBECONFIG environment variable.
  3. If you deployed dashboard-metrics-scraper in your cluster, set its endpoint to K8S_DASHBOARD_SIDECAR_HOST environment variable.
  4. Change directory into your dashboard source directory.
  5. Run hack/develop/run-dev-container.sh.

These manipulations will build container and run dashboard with make run as default.

To accessing Kubernetes Dashboard, open https://localhost:4443 from your browser.

Just to run development container without building and running dashboard

  1. Set K8S_DASHBOARD_CMD environment variable as bash.
  2. Run hack/develop/run-dev-container.sh.
  3. Run commands as you like in the container.

This runs container with bash command.

To run dashboard, execute make run. This will build dashboard for production and run three containers for the dashboard. Then, access https://localhost:4443 from your browser.

To access console inside of running development container

  1. Run docker exec -it k8s-dashboard-dev gosu user bash.

Copyright 2019 The Kubernetes Dashboard Authors