artemiscloud / activemq-artemis-operator-test-suite

activemq-artemis-operator-test-suite

The ActiveMQ Artemis Operator Test Suite.

Quick Start

Install Ginkgo, see https://onsi.github.io/ginkgo/

go install github.com/onsi/ginkgo/[email protected]

Set up a kubernetes cluster, see https://kind.sigs.k8s.io/docs/user/quick-start/

kind create cluster

Export KUBECONFIG, see https://kubernetes.io/docs/concepts/configuration/organize-cluster-access-kubeconfig/

export KUBECONFIG=$HOME/.kube/config

Clone activemq-artemis-operator, see https://github.com/artemiscloud/activemq-artemis-operator

git clone https://github.com/artemiscloud/activemq-artemis-operator.git

Use artemiscloud container images

Exexute the test suite using artemiscloud container images from quay.io and the local clone of the activemq-artemis-operator repository:

ginkgo -r -keepGoing test/smoke/basic/... -- \
        -broker-name activemq-artemis \
        -broker-image quay.io/artemiscloud/activemq-artemis-broker-kubernetes:dev.latest \
        -broker-image-second quay.io/artemiscloud/activemq-artemis-broker-kubernetes:dev.latest \
        -operator-image quay.io/artemiscloud/activemq-artemis-operator:dev.latest \
        -repository ../activemq-artemis-operator/deploy \
        -v2 -debug-run

Use local container images

The default operator image pull policy is Always, change it to use a local operator container image, see https://kubernetes.io/docs/concepts/containers/images/#updating-images

sed -i 's/imagePullPolicy:.*/imagePullPolicy: IfNotPresent/' ../activemq-artemis-operator/deploy/operator.yaml

Load local container images into the kubernetes cluster, see https://kind.sigs.k8s.io/docs/user/quick-start/#loading-an-image-into-your-cluster

kind load docker-image activemq-artemis-broker-kubernetes:dev.latest

Exexute the test suite using local container images and the local clone of activemq-artemis-operator repository

ginkgo -r -keepGoing test/smoke/basic/... -- \
        -broker-name activemq-artemis \
        -broker-image activemq-artemis-broker-kubernetes:dev.latest \
        -broker-image-second activemq-artemis-broker-kubernetes:dev.latest \
        -operator-image activemq-artemis-operator:dev.latest \
        -repository ../activemq-artemis-operator/deploy \
        -v2 -debug-run

Delete test suite namespaces

kubectl get namespaces | grep -o '^e2e[^ ]*' | while read NAMESPACE; do kubectl delete namespaces $NAMESPACE; done

Troubleshooting

Use the following command to investigate on kubernetes cluster failures:

• kubectl get pods --all-namespaces
• kubectl describe pod
• kubectl logs

activemq-artemis-operator-test-suite's Issues

Update hardcoded image to non-hardcoded one in upgrade test

Problem: x86-only image is hardcoded here:

activemq-artemis-operator-test-suite/test/smoke/basic/basic_update_test.go

Line 13 in d9f278b

CustomImage = "registry.redhat.io/amq7/amq-broker-rhel8@sha256:c5f4c08e068b9721967cf7c7cbd9a9e93fb5e39b264dd13b653e99d8f3fa9e0e"

solution:

a ) parse operator.yaml and extract relevant version for your architecture from it

b) add parameter to specify other image

Crash when trying to apply unsupported CRDs to the cluster (previous CRD versions to ocp 4.9)

Unsure what broke, just a note for later

Update -v2 switch properly to represent current state of lts vs non-lts