Development Guide

This document is intended to be the canonical source of truth for things like supported toolchain versions for building Voyager. If you find a requirement that this doc does not capture, please submit an issue on github.

This document is intended to be relative to the branch in which it is found. It is guaranteed that requirements will change over time for the development branch, but release branches of Voyager should not change.

Build Voyager

Some of the Voyager development helper scripts rely on a fairly up-to-date GNU tools environment, so most recent Linux distros should work just fine out-of-the-box.

Setup GO

Voyager is written in Google’s GO programming language. Currently, Voyager is developed and tested on go 1.9.2. If you haven’t set up a GO development environment, please follow these instructions to install GO.

Download Source

$ go get -u -v github.com/appscode/voyager
$ cd $(go env GOPATH)/src/github.com/appscode/voyager

Install Dev tools

To install various dev tools for Voyager, run the following command:

$ ./hack/builddeps.sh

Updating Codes

voyager usages codecgen to generate codes related to kubernetes. If changes happens to api types, codes needs to be regenerated. API types needs to be updated in both apis/voyager/v1beta1 and apis/voyager. Run the following command to generate codes:

$ ./hack/codegen.sh

Build Binary

$ ./hack/make.py
$ voyager version

Run Binary Locally

$ voyager run \
  --cloud-provider=minikube \
  --secure-port=8443 \
  --kubeconfig="$HOME/.kube/config" \
  --authorization-kubeconfig="$HOME/.kube/config" \
  --authentication-kubeconfig="$HOME/.kube/config" \
  --authentication-skip-lookup

Dependency management

Voyager uses Glide to manage dependencies. Dependencies are already checked in the vendor folder. If you want to update/add dependencies, run:

$ glide slow

Build Docker images

To build and push your custom Docker image, follow the steps below. To release a new version of Voyager, please follow the release guide.

# Build Docker image
$ ./hack/docker/voyager/setup.sh

# Add docker tag for your repository
$ docker tag appscode/voyager:<tag> <image>:<tag>

# Push Image
$ docker push <image>:<tag>

Build HAProxy

We package HAProxy and Kloader into a Ubuntu 16.04 based Docker image.

$ ./hack/docker/haproxy/<version>/setup.sh
$ ./hack/docker/haproxy/<version>/setup.sh push

Generate CLI Reference Docs

$ ./hack/gendocs/make.sh

Run Test

Run Short Unit Test by running

go test ./pkg/...

Run Full Test

To Run Full unit test You need to provide some secret in hack/configs/.env file. Or You may add them as environment variables.

TEST_GCE_SERVICE_ACCOUNT_DATA
TEST_GCE_PROJECT
TEST_ACME_USER_EMAIL
TEST_DNS_DOMAINS

Then run

$ ./hack/make.py test unit

Run e2e Test

To run e2e tests in minikube, add the following line to your machine’s /etc/hosts file:

$ minikube delete; minikube start
$ echo "$(minikube ip)   http.appscode.test" >> /etc/hosts

$ ./hack/docker/voyager/setup.sh; ./hack/docker/voyager/setup.sh push
$ env APPSCODE_ENV=dev ./hack/deploy/voyager.sh --provider=minikube
$ ./hack/make.py test minikube --selfhosted-operator
$ ./hack/make.py test minikube # Run Test against minikube, this requires minikube to be set up and started.

$ ./hack/make.py test e2e -cloud-provider=gce # Test e2e against gce cluster

$ ./hack/make.py test integration -cloud-provider=gce # Run Integration test against gce
                                                      # This requires voyager to be deployed in the cluster.

- Run only matching tests e2e test
$ ./hack/make.py test e2e -cloud-provider=gce -ginkgo.focus=<regexp>


- Run tests but do not delete resource that are created
$ ./hack/make.py test minikube -cloud-provider=gce -cleanup=false


- Run Service IP Persist test with provided IP
$ ./hack/make.py test e2e -cloud-provider=gce -lb-ip=35.184.104.215

Tests are run only in namespaces prefixed with test-. So, to run tests in your desired namespace, follow these steps:

# create a Kubernetes namespace in minikube with
kubectl create ns test-<any-name-you-want>

# run tests
./hack/make.py test minikube -namespace test-<any-name-you-want>

Full Spectrum of test configs

Following configurations can be enabled for test via flags in ./hack/make.py test.

Flag NameDefaultDescription
cloud-providerName of cloud Provider
ingress-class
namespacetest- Run tests in this namespaces
haproxy-imageappscode/haproxy:1.9.2-9.0.0-alpineHAProxy image name to run
cleanuptrueTurn off cleanup for dynamically generated pods and configmaps. Helps with manual testing
in-clusterfalseOperator is running inside cluster. Helps with running operator testing.
daemon-host-namemasterDaemon host name to run daemon hosts
lb-ipCheck load balancer IP with Static IP addressLoadBalancer persistent IP
rbacfalseCluster have RBAC enabled.
certfalseRun tests regarding certificates
dumpos.TempDir()Dump all Certificates and CA files for TLS ingress tests

e2e tests are powered by ginkgo. All the configs and flags of ginkgo are also available.

CRDs

voyager uses on two Custom Resource Definition object ingress.voyager.appscode.com and certificate.voyager.appscode.com. Those two objects can be created using the following command:

# Create Third Party Resources
$ kubectl apply -f https://raw.githubusercontent.com/appscode/voyager/9.0.0/apis/voyager/v1beta1/crds.yaml