

Connection Instruction between Tableau and FusionInsight¶

Succeeded Case¶

Tableau 10.0.0 ↔ FusionInsight HD V100R002C30 (Hive/SparkSQL)

Tableau 10.0.0 ↔ FusionInsight HD V100R002C50 (Hive/SparkSQL)

Tableau 10.1.4 ↔ FusionInsight HD V100R002C60U20 (Hive/SparkSQL)

Tableau 10.3.2 ↔ FusionInsight HD V100R002C70SPC200 (Hive/SparkSQL)

Tableau 10.5.0 ↔ FusionInsight HD V100R002C80SPC100 (Hive/SparkSQL)

Configure the Kerberos on Windows¶

Download and install MIT Kerberos from the following URL http://web.mit.edu/kerberos/dist/#kfw-4.0
Make sure the time differences between FusionInsight clusters and Tableau client is no longer than 5 minutes.
Configure required Kerberos filesystem

Created a user with "Human-Machine" as its type( For detail, take product documentation as a reference ), grant the Hive privileges to this user. For example, create a user named developuser, download the user.keytab and krb5.conf files on the Tableau client, rename the krb5.conf file into krb5.ini and save it into the following directory C:\ProgramData\MIT\Kerberos5

Configure Kerberos cache file
Create a directory to save the Kerberos cache file, for example, C:\temp
Configure the Environment Variables, Variable name is KRB5CCNAME, Variable value is C:\temp\krb5cache
restart the Tableau client
Start Kerberos Authentication on Windows
use the created username and password to log in, the Principal is equal to username@Kerberos Realm name
Open MIT Kerberos, click on Get Ticket, and type in the right Principal and Password for authentication

Connecting Tableau to Hive¶

Configure the ODBC interface to connect FusionInsight HiveDriver

Download and install the ODBC driver

Download URL: http://www.cloudera.com/content/cloudera/en/downloads/connectors/hive/odbc/hive-odbc-v2-5-15.html

and choose the right one depends on the OS and bit version

Configure ODBC drivers

Open ODBC Data Sources(64-bit) by searching the keyword ODBC on Windows

click on User DSN tab, click on Add button, choose Cloudera ODBC Driver for Apache Hive and click on Finish to start to configure
```
In detail:
1: Hive Server 2
2: No Service Discovery
3: 172.21.3.101
4: 21006
5: default
6: Kerberos
7: hadoop.hadoop.com
8: hive
9: SASL
```
click on Test button to test the connection
Open Tableau

Click on More option, and choose ODBC by search in keyword

Connection Configuration shown as bellow:

click on Connect and then click on Sign In

Search the Data
Search the data from multiple tables

Connecting Tableau to Spark¶

Download and install the ODBC driver for Spark Download urlhttp://www.tableau.com/support/drivers

Created DSN （Data Source Name）

Open ODBC Data Sources(64-bit)

Click on System DSN tab, click on Add, choose Simba Spark ODBC Driver and click on Finish

Open the installed Driver directory, for example, C:\Program Files\Simba Spark ODBC Driver\lib and open the DriverConfiguration64.exe to Configure

In detail:
1：SparkThriftServer (Spark 1.1 and later)
2: Kerberos
3: hadoop.hadoop.com
4: spark
5: SASL

click on Advanced Options and choose "Driver Config Take Precedence"

Click on ok to save the configuration

Open Tableau

Click on More option, and choose Spark SQL by search in keyword

Connection Configuration shown as bellow:

Server info can be got from FusionInsight Manager Web UI

Port info can be got from FusionInsight Manager Web UI as well

Click on Sign In , to come into a new Page, choose Schema and Table shown as bellow

Open Sheet to visulize the data
Performance test
Search the table web_sales which contains millions of records
Search the table by multiple tables whose names are store_sales and item

Add customer_address table

Test outcome：

FAQ¶

Cannot find C:\ProgramData\MIT\Kerberos5

This Folder is hidden, configure the windows can solve it

Connection succeeded but permission denied

Use the user who has the privilege to DATABASE