task.yml

- name: Create new Elastic Search  load balancer if it doesn't exists
  ec2_elb_lb:
    name: "{{ elb_name }}"
    state: present
    region: "{{aws_region}}"
    scheme: internal
    subnets:
      - "{{ elb_subnet_1 }}"
      - "{{ elb_subnet_2 }}"
      - "{{ elb_subnet_3 }}"
    security_group_ids: "{{vpc_sg}}"
    listeners:
      - protocol: tcp
      - load_balancer_port: 9600
      - instance_port: 9600    
    stickiness:
      type: loadbalancer
      enabled: yes
      expiration: 300
    access_logs:
      interval: 5 # minutes (defaults to 60)
      s3_location: "{{ lb_log_location }}" # This value is required if access_logs is set
      s3_prefix: "logstash"
    cross_az_load_balancing: "yes"
  # when:  ' "{{lb_name}}" != "{{ elb_name }}" '
  when: elb_name_output is undefined

Hi @farahfa. Does this problem persist if you specify listeners like this?

listeners:
  - protocol: tcp
    load_balancer_port: 9600
    instance_port: 9600

Right now your playbook reads as though it is designating three listeners rather than one listener with those three attributes.