run-rabbitmq

Requirements:

• kubectl
• Operator Lifecycle Manager
• yq

That last one is more of a recommendation than a hard requirement, but anyway.

Get an appropriate kubeconfig file, then switch to appropriate Kubernetes context. Here, our imaginary Kubernetes cluster is named "yoyodyne-ops":

kubectl config get-contexts
kubectl config use-context yoyodyne-ops
kubectl config get-contexts
kubectl cluster-info
kubectl get all -A

Install OLM:

operator-sdk olm install --version v0.28.0
operator-sdk olm status
kubectl events -A -w

Create a Namespace, and install the operator:

kubectl create namespace rabbitmq
kubectl apply -f operatorgroup.yaml
kubectl apply -f subscription.yaml

At this point it should be mentioned that this kind of "manual" approach to installation of anything is likely not how things should look like in environments even remotely resembling "production".

Look at those Kubernetes events, then get the RabbitMQ installation going, and watch more events:

kubectl events -A -w
kubectl apply -f rabbitmq.yaml
kubectl events -A -w

Check out the Pods and follow RabbitMQ's logs:

kubectl get pods -n rabbitmq
kubectl logs -n rabbitmq -l app.kubernetes.io/name=rabbitmq --follow

Fish out the credentials for RabbitMQ's default user, then forward local ports to RabbitMQ proper and its web management console, through RabbitMQ's service:

kubectl get secret -n rabbitmq rabbitmq-default-user -o yaml | yq '.data.username' | base64 -d
kubectl get secret -n rabbitmq rabbitmq-default-user -o yaml | yq '.data.password' | base64 -d
kubectl port-forward -n rabbitmq service/rabbitmq 5672:5672 15672:15672

Then for example access the web management console by pointing your browser to:

http://localhost:15672/

Note: with those kubectl port-forwards in place, RabbitMQ clients should talk either to localhost:5672, in case they can access the host's network directly, or to host.docker.internal:5672 in case they're running in a Docker container.