Configuring Ambari for Non-Root user

[lhoss]

Impl. for feature request #119
By default the server and agent are (still) run as root user.

To activate the new feature, and run both services under a new user ambari (deployment of the OS user is also done, incl. the extensive sudoers rules required), add following group vars

ambari_server_user: ambari
ambari_agent_user: ambari

I successfully tested with a deployment with following config (overrides vs the default all) on one 32GB AWS node :

ambari_version: '2.7.3.0'                                 # must be the 4-part full version number
hdp_version: '3.1.0.0'                                    # must be the 4-part full version number
hdf_version: '3.3.1.0'                                    # must be the 4-part full version number
hdp_build_number: 'auto'                                  # the HDP build number from docs.hortonworks.com (if set to 'auto', Ansible will try to get it from the repository)
hdf_build_number: 'auto'                                  # the HDF build number from docs.hortonworks.com (if set to 'auto', Ansible will try to get it from the repository)
hdpsearch_version: '4.0.0'                                # must be the full version number
hdpsearch_build_number: '400'                             # the HDP Search build number from docs.hortonworks.com (hardcoded to 100 for the moment)

java: 'openjdk'                                          # can be set to 'embedded', 'openjdk' or 'oraclejdk'

database: 'postgres'                                      # can be set to 'embedded', 'postgres', 'mysql' or 'mariadb'

#security: 'mit-kdc'                                       # can be set to 'none', 'mit-kdc' or 'active-directory'
security: none

#############################
## blueprint configuration ##
#############################

blueprint_name: '{{ cluster_name }}_blueprint'            # the name of the blueprint as it will be stored in Ambari
blueprint_file: 'blueprint_dynamic.j2'                    # the blueprint JSON file - 'blueprint_dynamic.j2' is a Jinja2 template that generates the required JSON

blueprint_dynamic_clients:
      - 'ZOOKEEPER_CLIENT'
      - 'HDFS_CLIENT'
      - 'YARN_CLIENT'
      - 'MAPREDUCE2_CLIENT'
      - 'TEZ_CLIENT'
      - 'PIG'
      - 'SQOOP'
      - 'HIVE_CLIENT'
      - 'OOZIE_CLIENT'
      - 'INFRA_SOLR_CLIENT'
      - 'SPARK2_CLIENT'
      - 'HBASE_CLIENT'
      - 'ATLAS_CLIENT'

blueprint_dynamic:                                        # properties for the dynamic blueprint - these are only used by the 'blueprint_dynamic.j2' template to generate the JSON
  - host_group: "hdp-master"
    clients: "{{ blueprint_dynamic_clients }}"
    services:
      - ZOOKEEPER_SERVER
      - NAMENODE
      - SECONDARY_NAMENODE
      - RESOURCEMANAGER
      - APP_TIMELINE_SERVER
      - YARN_REGISTRY_DNS
      - TIMELINE_READER
      - HBASE_MASTER
      - HIVE_SERVER
      - HIVE_METASTORE
      - OOZIE_SERVER
      - HISTORYSERVER
      - SPARK2_JOBHISTORYSERVER
      - RANGER_ADMIN
      - ATLAS_SERVER
      - KNOX_GATEWAY
      - ZEPPELIN_MASTER
      - AMBARI_SERVER
      - INFRA_SOLR
      - HST_SERVER
      - ACTIVITY_ANALYZER
      - ACTIVITY_EXPLORER
      - HST_AGENT
      - METRICS_COLLECTOR
      - METRICS_GRAFANA
      - METRICS_MONITOR
      - REGISTRY_SERVER
      - NIFI_REGISTRY_MASTER
      - RANGER_KMS_SERVER
      - RANGER_USERSYNC
  # - host_group: "hdp-slave"
  #   clients: "{{ blueprint_dynamic_clients }}"
  #   services:
      - DATANODE
      - NODEMANAGER
      - METRICS_MONITOR
      - HST_AGENT
      - HBASE_REGIONSERVER
      - NIFI_MASTER
      - KAFKA_BROKER
      # NEW test SOLR (this is not the same as INFRA_SOLR)
      - SOLR_SERVER

ps: Initial test with ambari_version: '2.7.4.0' failed , see bug #166

[lhoss]

@alexandruanghel this new feature is now done and tested.
Let me know if/where (in the README) we can document the 2 new variables

[lhoss]

@alexandruanghel this new feature is now better tested (incl. many deployments at a customer), after I added some fixes