Example: Hive Impersonation

The following examples illustrate Hive impersonation:

Example 1

  1. Log in as a non-mapr user and generate a Data Fabric ticket:
    $ su mapruser1
$ maprlogin password
  2. Connect to HiveServer2:
    
$ hive --service beeline
Beeline version 2.3.3-mapr-SNAPSHOT by Apache Hive
beeline> !connect jdbc:hive2://node4.cluster.com:10000/default;ssl=true;auth=maprsasl
Connecting to jdbc:hive2://node4.cluster.com:10000/default;ssl=true;auth=maprsasl
Connected to: Apache Hive (version 2.3.3-mapr-SNAPSHOT)
Driver: Hive JDBC (version 2.3.3-mapr-SNAPSHOT)
Transaction isolation: TRANSACTION_REPEATABLE_READ
  3. Create a table, and upload data:
    
0: jdbc:hive2://node4.cluster.com:10000/defau> create table impersonation_example_first (id int, username string);
0: jdbc:hive2://node4.cluster.com:10000/defau> insert into impersonation_example_first values (1, 'mapruser1');
  4. To check that impersonation works, use the following commands to check the /warehouse directory of the Data Fabric file system:
    
$ hadoop fs -ls /user/hive/warehouse
Found 1 items
drwxr-xr-x   - mapruser1 mapruser1          1 2019-05-22 14:40 
/user/hive/warehouse/impersonation_example_first

$ hadoop fs -ls /user/hive/warehouse/impersonation_example_first
Found 1 items
-rwxrwxrwx   3 mapruser1 mapruser1         12 2019-10-15 07:21 
/user/hive/warehouse/impersonation_example_first/000000_0

Example 2

  1. Generate a Data Fabric ticket for a non-mapr user.
  2. Connect through JDBC using the hive.server2.proxy.user option with a non-mapr user name as an argument:
    $ hive --service beeline
beeline> !connect 
jdbc:hive2://node4.cluster.com:10000/default;auth=maprsasl;ssl=true;hive.server2.proxy.user=mapruser1
Connecting to 
jdbc:hive2://node4.cluster.com:10000/default;auth=maprsasl;ssl=true;hive.server2.proxy.user=mapruser1
Client: auth-conf,auth-int,auth.Using Server one
Connected to: Apache Hive (version 2.3.3-mapr-SNAPSHOT)
Driver: Hive JDBC (version 2.3.3-mapr-SNAPSHOT)
Transaction isolation: TRANSACTION_REPEATABLE_READ
  3. Create a table and upload data:
    0: jdbc:hive2://node4.cluster.com:10000/default> create table impersonation_example_second (id int);
0: jdbc:hive2://node4.cluster.com:10000/default> insert into table impersonation_example_second values (1), (2), (3), (5);
  4. Check the owner of the table and data:
    $ hadoop fs -ls /user/hive/warehouse/impersonation_example_second
Found 1 items
drwxrwxrwx   - mapruser1 mapruser1          1 2019-05-23 12:29 
/user/hive/warehouse/impersonation_example_second

$ hadoop fs -ls /user/hive/warehouse/impersonation_example_second
Found 1 items
-rwxrwxrwx   3 mapruser1 mapruser1          8 2019-05-23 12:29 
/user/hive/warehouse/impersonation_example_second/000000_0