@Manoj Dhake - Depending on how Atlas was installed, you may need to configure the Atlas post execution Hive hook (see: https://docs.hortonworks.com/HDPDocuments/HDP2/HDP-2.3.0/bk_installing_manually_book/content/configuring_atlas.html). Once this is completed, Atlas will receive notifications from hive for all queries & DDL statements executed resulting in entity creations in Atlas which can be subsequently searched and viewed through the UI. To import all existing Hive databases, tables, views etc., you can execute the hive-bridge which will perform an import of all of the existing data in the catalog ($ATLAS_HOME/bin/import-hive.sh). This is typically a one time operation. Once the above steps are completed, you will be able to search for entities in Atlas using the Atlas UI. Atlas 0.5 (included in HDP-2.3/2.4 provides integration with Hive alone. Atlas 0.7, to be released with HDP-2.5, includes integrations with a broader set of services, including Sqoop, Falcon, Storm & improved Hive integration as well as a slew of other features such as HA, Ranger integration, security, UI improvements, Business Taxonomy etc. ... View more

Hi @Andrew Watson, No such script exists today, although one could be written for the specific version of Hue in use as well as the backing database. In the upcoming release of Ambari, an Ambari view has been introduced to allow for just such a migration (https://github.com/apache/ambari/tree/trunk/contrib/views/hueambarimigration). Until such time that this is released, a custom solution would need to be implemented. ... View more

@Vasilis Vagias - In the case of a CTAS operation in hive, tags in Atlas that are associated with the source entities (columns, tables, views etc.) are not propagated to resulting entities in Atlas. This is mostly a function of Atlas not currently providing column level lineage. Both column level lineage as well as tag propagation are roadmap items. In short, the behavior that you are seeing is expected. ... View more

@wayne2chicago Ambari will only install Atlas if you select during install (UI install wizard) or if you are using blueprints, you explicitly declare that Atlas should be installed. Ranger does not have a dependency on Atlas. As for the Hive query that is failing, in HDP 2.3x, Atlas is required to be running in order for the Hive post execution hook for Atlas to fire properly since it is using a REST API to propagate the metadata. This limitation is being addressed by introducing a messaging layer which decouples Atlas from Hive such that the state of the Atlas process does not affect the execution of Hive queries. ... View more

@Jagdish Saripella - Atlas 0.5 does not support authentication besides SPNEGO/Kerberos. Proper authentication is planned for 0.7 release and is captured here: https://issues.apache.org/jira/browse/ATLAS-494. ... View more

@Pardeep Gorla - If import-hive.sh was executed correctly, then you should be able to query for hive tables through the Atlas UI using syntax such as: hive_table hive_table where tableName="employee" This will make use of the Atlas DSL to search as opposed to the free text search which the Atlas UI will default to IF the DSL query does not yield any results. ... View more

Starting in HDP 2.3, Atlas provides mechanisms that allow for automatic creation of hive entities (DBs, Tables, Views, Columns) through a hive post execution hook. These entities can then be decorated with additional metadata through the association of Traits (think complex attribute bearing Tags). Atlas also provides an extensible mechanism which allows for the creation of a business term related taxonomy. This separate taxonomy can then be associated with specific Hive assets (entities). For more information, refer to http://incubator.apache.org/projects/atlas.html. ... View more