[2324307783] Upstream Reporter: Armel Soro
Upstream issue status: Open
Upstream description:

/kind bug

Follow-up issue discovered while working on https://github.com/janus-idp/operator/pull/369

What versions of software are you using?

• Operator on current main branch: https://github.com/janus-idp/operator/commit/5b0aeaa23beaddc2d5810c7aea554d763c0e0b96

What did you run exactly?

• Run the operator against a cluster

make install run

• In a separate terminal, apply any of the examples, e.g. examples/bs1.yaml:

kubectl apply -f examples/bs1.yaml

• Inspect the local DB StatefulSet and the Services created by the operator

kubectl get statefulset --selector=app.kubernetes.io/name=backstage,app.kubernetes.io/instance=bs1 -o yaml | yq '.items[0].spec'

kubectl get service --selector=app.kubernetes.io/name=backstage,app.kubernetes.io/instance=bs1

Actual behavior

• Here are the services created by the operator:

$ kubectl get service --selector=app.kubernetes.io/name=backstage,app.kubernetes.io/instance=bs1                              
NAME               TYPE        CLUSTER-IP      EXTERNAL-IP   PORT(S)    AGE
backstage-bs1      ClusterIP   172.31.39.19    <none>        80/TCP     6m36s
backstage-db-bs1   ClusterIP   172.31.20.189   <none>        5432/TCP   6m36s

• And the spec of the DB StatefulSet, where we can see that spec.serviceName is hardcoded to backstage-psql-cr1-hl (and spec.template.metadata.name is also hardcoded to backstage-db-cr1, but this might not be an issue per se).

persistentVolumeClaimRetentionPolicy:   whenDeleted: Retain
  whenScaled: Retain
podManagementPolicy: OrderedReady
replicas: 1
revisionHistoryLimit: 10
selector:   matchLabels:     rhdh.redhat.com/app: backstage-db-bs1
serviceName: backstage-psql-cr1-hl
template:   metadata:     creationTimestamp: null
    labels:       rhdh.redhat.com/app: backstage-db-bs1
    name: backstage-db-cr1
  spec:     automountServiceAccountToken: false
    containers:       - env:           - name: POSTGRESQL_PORT_NUMBER
            value: "5432"
          - name: POSTGRESQL_VOLUME_DIR
            value: /var/lib/pgsql/data
          - name: PGDATA
            value: /var/lib/pgsql/data/userdata
        envFrom:           - secretRef:               name: backstage-db-bs1
        image: quay.io/fedora/postgresql-15:latest
        imagePullPolicy: IfNotPresent
        livenessProbe:           exec:             command:               - /bin/sh
              - -c
              - exec pg_isready -U ${POSTGRES_USER} -h 127.0.0.1 -p 5432
          failureThreshold: 6
          initialDelaySeconds: 30
          periodSeconds: 10
          successThreshold: 1
          timeoutSeconds: 5
        name: postgresql
        ports:           - containerPort: 5432
            name: tcp-postgresql
            protocol: TCP
        readinessProbe:           exec:             command:               - /bin/sh
              - -c
              - -e
              - |
                exec pg_isready -U ${POSTGRES_USER} -h 127.0.0.1 -p 5432
          failureThreshold: 6
          initialDelaySeconds: 5
          periodSeconds: 10
          successThreshold: 1
          timeoutSeconds: 5
        resources:           limits:             cpu: 250m
            ephemeral-storage: 20Mi
            memory: 1Gi
          requests:             cpu: 250m
            memory: 256Mi
        securityContext:           allowPrivilegeEscalation: false
          capabilities:             drop:               - ALL
          runAsGroup: 0
          runAsNonRoot: true
          seccompProfile:             type: RuntimeDefault
        terminationMessagePath: /dev/termination-log
        terminationMessagePolicy: File
        volumeMounts:           - mountPath: /dev/shm
            name: dshm
          - mountPath: /var/lib/pgsql/data
            name: data
    dnsPolicy: ClusterFirst
    restartPolicy: Always
    schedulerName: default-scheduler
    securityContext: {}
    serviceAccount: default
    serviceAccountName: default
    terminationGracePeriodSeconds: 30
    volumes:       - emptyDir:           medium: Memory
        name: dshm
updateStrategy:   rollingUpdate:     partition: 0
  type: RollingUpdate
volumeClaimTemplates:   - apiVersion: v1
    kind: PersistentVolumeClaim
    metadata:       creationTimestamp: null
      name: data
    spec:       accessModes:         - ReadWriteOnce
      resources:         requests:           storage: 1Gi
      volumeMode: Filesystem
    status:       phase: Pending

Expected behavior

Per the official K8s docs, a Headless Service is required for StatefulSets and must exist:

StatefulSets currently require a Headless Service to be responsible for the network identity of the Pods. You are responsible for creating this Service

Not sure why K8s does not return an error with this non-existing service, but the expected behavior would be:

• spec.serviceName should not be hardcoded. Instead, it can have a value like backstage-psql-<cr_name>-hl (and this headless service should be created by the operator as well)
• spec.template.metadata.name should not be hardcoded. Instead, it can have a value like backstage-psql-<cr_name>

Any logs, error output, etc?

None