spark-hadoop.yml

---
#
# Install basic packages and python
#
- name: install basic packages, python, and thunder
  hosts: all
  become: yes
  become_user: root
  
  vars_files: 
    - user_data/user-defines.yml
    - user_data/user-passwords.yml 
    
  vars:
    ssh_public_key_content: "{{ lookup('file', ssh_public_key ) }}"
   
  pre_tasks:
    - name: Update APT cache
      apt: update_cache=yes

  roles: 
    - common
    - {role: supervisord, tags: supervisor}
    - {role: miniconda, tags: miniconda}
    - {role: thunder, 
       tags: thunder, 
       environment: {PATH: "/usr/local/miniconda/bin:{{ ansible_env.PATH }}"}}

#
# Starting up spark, hadoop, and jupyterhub
# 
# Note: we do this after the python installation because we want to 
# use the custom python installed above instead of system python
#
- name: install and deploy spark and hdfs
  hosts: all
  become: yes
  become_user: root
  environment:
    PATH: "/usr/local/miniconda/bin:{{ ansible_env.PATH }}"
    PYSPARK_PYTHON: "/usr/local/miniconda/bin/python"
  vars_files: 
    - user_data/user-defines.yml

  tags: spark-hadoop

  roles:
    - {role: spark, tags: spark}   
    - {role: hadoop, tags: hadoop}
    
- name: deploy jupyterhub
  hosts: spark_masters
  become: yes
  become_user: root
  environment:
    PATH: "/usr/local/miniconda/envs/python35/bin:{{ ansible_env.PATH }}"
  tags: jupyterhub

  vars_files: 
    - user_data/user-defines.yml
    - user_data/user-passwords.yml

  roles:
    - jupyterhub