Deleting a workload won't delete the component and http-server config

[Mees-Molenaar]

Hi,

while testing the runtime-operator I think I found a bug.
It seems that when deleting a workload, you still can call the component in that workload via the http-server.

Steps to reproduce:

1. Run the runtime-operator
   a. make kind-setup
   b. make helm-install
2. Applied a workload: kubectl apply -f workload.yaml
3. Performed a curl request: curl -v -H "Host: localhost" http://localhost:8888/
4. Got the expected response
5. Deleted the workload: kubectl delete -f workload.yaml
6. Verified it by: kubectl get workloads
7. Performed the same curl request: curl -v -H "Host: localhost" http://localhost:8888/
8. Got the same response as in 4, I would expect that it doesn't resolve anymore

I am not sure how a runtime-operator works, but I would expect that the component was deleted and the http-server config host will be removed.

Workload yaml:

apiVersion: runtime.wasmcloud.dev/v1alpha1
kind: WorkloadDeployment
metadata:
  name: long-running-component
spec:
  replicas: 1
  template:
    spec:
      hostSelector:
        hostgroup: public-ingress
      components:
        - name: long-running-component-2
          image: kind-registry.default:5001/long-running-component:0.1.1
      hostInterfaces:
        - namespace: wasi
          package: http
          interfaces:
            - incoming-handler
          config:
            host: "localhost"

kind-config.yaml

kind: Cluster
apiVersion: kind.x-k8s.io/v1alpha4
nodes:
  - role: control-plane
    extraPortMappings:
      - containerPort: 30950
        hostPort: 8888
        protocol: TCP
containerdConfigPatches:
  - |-
    [plugins."io.containerd.grpc.v1.cri".registry.mirrors."localhost:5001"]
      endpoint = ["http://kind-registry:5000"]

Kind regards,
Mees

[lxfontes]

thank you! noticed this too and I think it is on the host side.
we are starting our pre-release checks and will address soon