Running the End-To-End Tests

The following environment variables should be set:

1. E2E_NAMESPACE is the namespace where the tests should be run.
2. CLUSTER_DOMAIN is the DNS domain managed in which ingresses should be created
3. CLUSTER_DOMAIN_INTERNAL is the internal DNS domain managed in which ingresses should be created
4. CONTROLLER_ID is set so that all stacks are only managed by the controller being currently tested.
5. KUBECONFIG with the path to the kubeconfig file

To run the tests run the command:

go test -parallel $NUM_PARALLEL github.com/zalando-incubator/stackset-controller/cmd/e2e

Over here $NUM_PARALLEL can be set to a sufficiently high value which indicates how many of the parallel type tests can be run concurrently.

Example - run E2E test

1. Start apiserver proxy

kubectl proxy

2. use kubectl watch to show what happens

watch -n 10 "kubectl get -n foo stackset,stack,ing,ep,deployment"

3. recreate namespace foo and run local build stackset-controller

kubectl delete namespace foo; kubectl create namespace foo
make
./build/stackset-controller --apiserver=http://127.0.0.1:8001 \
--enable-configmap-support --enable-secret-support --enable-routegroup-support \
--enable-pcs-support --enable-traffic-segments --controller-id=foo \
--sync-ingress-annotation=example.org/i-haz-synchronize \
--sync-ingress-annotation=teapot.org/the-best \
--cluster-domain=${CLUSTER_DOMAIN} --cluster-domain=${CLUSTER_DOMAIN_INTERNAL}

4. rebuild e2e test and run e2e tests in foo namespace

rm -f build/e2e; make build/e2e
CLUSTER_DOMAIN=example.org CLUSTER_DOMAIN_INTERNAL=ingress.cluster.local CLUSTER_NAME=example E2E_NAMESPACE=foo CONTROLLER_ID=foo KUBECONFIG=$HOME/.kube/config ./build/e2e -test.v #-test.run=TestTrafficSwitch