Hi Team, I have been recently trying out the flink-k8s-operator for session mode for the auto-scaler feature. Based on the testing that we have done till date its found that - 1. We wont be able to submit an entirely different docker image to the existing cluster ( which runs on a base image) a. its because the kind FlinkSessionJob does not support any field or arg for passing the images 2. Now we were checking the possibility of passing a zipped virtual env, while submitting the job a. For this we were focusing on the argument -pyarch and were passing the blob path to the zipped env file b. The problem that we are facing here is - the flink system is not capturing the zipped env and the package that we have supplied within the new venv via the zip is said to not found

apiVersion: flink.apache.org/v1beta1
kind: FlinkDeployment
metadata:
  name: flink-session
  namespace: flink-test
spec:
  image: pyflink-session-test:v1.0
  flinkVersion: v1_20
  imagePullPolicy: Always
  serviceAccount: flink
  mode: standalone
  
  flinkConfiguration:
    # All Flink runtime config keys go here
    fs.allowed-fallback-filesystems: "file"
    io.tmp.dirs: "/tmp"
    taskmanager.numberOfTaskSlots: "4"
    
    # ===== OPERATOR AUTOSCALER =====
    kubernetes.operator.job.autoscaler.enabled: "true"
    kubernetes.operator.job.autoscaler.target.utilization: "0.7"
    kubernetes.operator.job.autoscaler.target.utilization.boundary: "0.2"
    kubernetes.operator.job.autoscaler.stabilization.interval: "1m"
    kubernetes.operator.job.autoscaler.metrics.window: "5m"
    kubernetes.operator.job.autoscaler.scale-up.grace-period: "1m"
    kubernetes.operator.job.autoscaler.scale-down.grace-period: "5m"
    
  # # 💡 MOVED: jobManager must be a direct child of 'spec'
  jobManager:
    replicas: 1
    resource:
      memory: "2048m"
      cpu: 1
  taskManager:
    replicas: 2
    resource:
      # memory: "10240m"
      memory: "2048m"
      cpu: 2

apiVersion: flink.apache.org/v1beta1
kind: FlinkSessionJob
metadata:
  name: ppe-kafka-streaming
  namespace: flink-test
spec:
  deploymentName: flink-session
  
  job:

    jarURI: /opt/flink/opt/flink-python-1.20.3.jar
    # entryClass: org.apache.flink.client.python.PythonDriver
    # args:
    #   # 1. Main PyFlink Script
    #   - "--python"  # Changed from -py
    #   - "/opt/flink/usrlib/streaming_test.py"
      
    #   # 2. Python Archives
    #   - "--pyArchives"  # Changed from -pyarch
    #   - "blob_path#flink_venv"
      
    #   # 3. Python Executable
    #   - "-pyexec"  # This is correct!
    #   - "flink_venv/bin/python" 

    args:
    - "-py"
    - "/opt/flink/usrlib/streaming_test.py"
    - "-pyarch"
    - "blob_path#flink_venv"
    - "-pyexec"
    - "flink_venv/bin/python"
    
    parallelism: 2

Request for your help if you have faced and tackled a similar issue. #C065944F9M2 #C03G7LJTS2G #C03GV7L3G2C