Contents:
This section describes how to ensure that the Designer Cloud Powered by Trifacta® platform is configured correctly to connect to Hive when Ranger is enabled for Hive. Ranger provides role-based authorization for Hive and other Hadoop components on the Hortonworks platform.
- This documentation was last validated for the Ranger version included with HDP 2.3.x.
Ranger effectively functions as a whitelist of URI's; by default, access is denied for any object in Hive. When a URI is requested, Ranger checks HDFS for permissions for the authenticated user. If HDFS denies access, then Ranger checks its defined set of URI's for the permission and, if a match is found, grants access for the authenticated user.
- For more information, see http://hortonworks.com/blog/best-practices-for-hive-authorization-using-apache-ranger-in-hdp-2-2/
Before you begin, please verify that your enterprise has deployed both Hive and Ranger according to recommended configuration practices. For more information, please consult the documentation that was provided with your Hadoop distribution. NOTE: Before you begin, you must integrate the Designer Cloud Powered by Trifacta platform with Hive. See Configure for Hive. Secure impersonation ensures consistent and easily traceable security access to the data stored within your Hadoop cluster. NOTE: Although not required, secure impersonation is highly recommended for connecting the platform with Hive. Since secure impersonation for the combination of Hive Server 2 and Ranger is not supported by Ranger, you must apply the following additional configuration changes to the Designer Cloud Powered by Trifacta platform to enable secure impersonation in the environment: You can apply this change through the Admin Settings Page (recommended) or Set the following parameter: When the Designer Cloud Powered by Trifacta platform is enabled with secure impersonation and submits requests to Hive, the following steps occur: The Hive server authorizes access to the underlying table through Ranger as the Hadoop principal user assigned to NOTE: This Hadoop principal is the user that should be configured through policies in Ranger to have the appropriate privileges. NOTE: Since Ranger assigns access to databases, tables, and columns to Unix users and groups, a common practice is to assign the Hadoop principal users (used by Alteryx users) to dedicated Unix groups that are separate from the Unix group NOTE: In UNIX environments, usernames and group names are case-sensitive. Please verify that you are using the case-sensitive names for users and groups in your Hadoop configuration and Alteryx configuration file. In Ranger, you can configure access through policies. A Ranger policy is a combination of: NOTE: In general, to manage access through Ranger, permissions to underlying Hadoop components such as HDFS or Hive should be minimized within those components. All permissions in Ranger are additive, which means that you should be careful about overlapping users and groups. The following components require these permissions at a minimum to be assigned to the Hadoop principal: Select, Update. Create (for Hive publishing)Pre-requisites
Secure Impersonation with Designer Cloud Powered by Trifacta platform and Hive with Ranger
hive to the Unix or LDAP group [os.group (default=trifacta)]. trifacta-conf.json.
For more information, see Platform Configuration Methods."hdfs.permissions.userUmask" = 027
Users and Groups for Ranger
[hadoop.user.principal (default=trifacta)] user through Kerberos.[hadoop.user.principal]. hive, which should be part of the group [hadoop.group (default=trifactausers)].[os.group (default=trifacta)] use within Ranger. Ranger should not grant any privileges and roles to the Unix group [os.group (default=trifacta)].Policies in Ranger
Required Permissions
Component Permissions HDFS Read, Write, Execute Hive
Verify Operations
After you have completed your configuration changes, you should restart the platform. See Start and Stop the Platform.
To verify platform operations, run a simple job. For more information, see Verify Operations.
This page has no comments.