setup_spark_distribution.yml

---
- name: Check if spark in installed
  stat: "path=/home/{{ ansible_user }}/spark-{{ spark_version }}-bin-hadoop3/bin"
  register: spark_installation_path

- name: Download spark distribution in master if not installed
  get_url:
    url: "https://dlcdn.apache.org/spark/spark-{{ spark_version }}/spark-{{ spark_version }}-bin-hadoop3.tgz"
    dest: "/home/{{ ansible_user }}/spark-{{ spark_version }}-bin-hadoop3.tgz"
    checksum: "sha512:{{ spark_sha512_checksum }}"
  delegate_to: master
  when: spark_installation_path.stat.isdir is not defined or not spark_installation_path.stat.isdir
  throttle: 1

- name: Sync spark from master to worker nodes
  ansible.posix.synchronize:
    src: "/home/{{ ansible_user }}/spark-{{ spark_version }}-bin-hadoop3.tgz"
    dest: "/home/{{ ansible_user }}/spark-{{ spark_version }}-bin-hadoop3.tgz"
    mode: push
  delegate_to: master
  when:
    - inventory_hostname != 'master'
    - spark_installation_path.stat.isdir is not defined or not spark_installation_path.stat.isdir
  become: no

- name: Decompress the spark distribution
  unarchive:
    src: "/home/{{ ansible_user }}/spark-{{ spark_version }}-bin-hadoop3.tgz"
    dest: "/home/{{ ansible_user }}"
    remote_src: yes
    creates: "/home/{{ ansible_user }}/spark-{{ spark_version }}-bin-hadoop3/bin"
  register: installation_task

- name: Delete the spark download if successfully installed
  file:
    path: "/home/{{ ansible_user }}/spark-{{ spark_version }}-bin-hadoop3.tgz"
    state: absent
  when: installation_task is succeeded or installation_task is skipped

- name: Symlink decompressed spark directory to spark home
  file:
    src: "/home/{{ ansible_user }}/spark-{{ spark_version }}-bin-hadoop3"
    dest: "{{ spark_home }}"
    state: link