Initial failed deployment can not be fixed

Affected version

0.5.0-nightly

Current and expected behavior

• stackablectl --namespace=stackable operator install airflow
  

• helm repo add bitnami https://charts.bitnami.com/bitnami
  

• helm install airflow-postgresql bitnami/postgresql --version 11.0.0 \
  --set auth.username=airflow \
  --set auth.password=airflow \
  --set auth.database=airflow
  

• helm install redis bitnami/redis \
  --set auth.password=redis
  

• Wait for all ressources are deployed

• create airflow file

• # airflow.yml
  apiVersion: v1
  kind: Secret
  metadata:
    name: simple-airflow-credentials
  type: Opaque
  stringData:
    adminUser.username: airflow
    adminUser.firstname: Airflow
    adminUser.lastname: Admin
    adminUser.email: airflow@airflow.com
    adminUser.password: airflow
    connections.secretKey: thisISaSECRET_1234
    connections.sqlalchemyDatabaseUri: postgresql+psycopg2://airflow:airflow@airflow-postgresql.WRONG-VALUE.svc.cluster.local/airflow
    connections.celeryResultBackend: db+postgresql://airflow:airflow@airflow-postgresql.WRONG-VALUE.svc.cluster.local/airflow
    connections.celeryBrokerUrl: redis://:redis@redis-master:6379/0
  
  ---
  apiVersion: airflow.stackable.tech/v1alpha1
  kind: AirflowCluster
  metadata:
    name: airflow
  spec:
    version: 2.2.4-python3.9-stackable0.3.0
    statsdExporterVersion: v0.22.4
    executor: CeleryExecutor
    loadExamples: true
    exposeConfig: false
    credentialsSecret: simple-airflow-credentials
    webservers:
      roleGroups:
        default:
          replicas: 1
    workers:
      roleGroups:
        default:
          replicas: 2
    schedulers:
      roleGroups:
        default:
          replicas: 1
  

•   kubectl apply -f airflow.yml
  

This config has a bug (wrong database url)

The operator now tries to deploy airflow but the database init fails.
After seven retries all the retries are done and we are in a failure state.
Now I deleted the deployment with
kubectl delete -f airflow.yml

After that i fixed the bug and run kubectl apply -f airflow.yml. This does not try again the database init and just goes to error state.

My expected behavior would be that the database init would tried again and succeeds this time.

Possible solution

Delete the airflow operator and the Custom ressource definition and reinstall the operator. Then apply the correct yaml file. This time the database init succeeds.

Additional context

No response

Environment

Client Version: v1.22.4
Server Version: v1.23.9

Would you like to work on fixing this bug?

No response

Use airflow.yml and the AirflowCluster reconciliation path as the entry point; reproduce the failed PostgreSQL initialization, delete and reapply the resource, and inspect how the operator records terminal failure after its retries. Done means correcting the connection and reapplying causes database initialization and cluster deployment to retry and succeed without reinstalling the operator or its CRD.