Data Connection to Apache Impala

Version

Application Scenario

This document introduces how to connect to the Apache Impala data source.

Note:

If multiple databases exist in Impala, only one database can be connected through one data connection.

Version/Driver

Download the driver package and upload it to FineBI. For details, see section "Uploading the Database Driver" in Driver Management.

Connection Information Collection

Before connecting the database, you need to collect the following information:

• IP address and port number of the database server

• Database name

• Username and password (if username and password authentication is used); client principal and keytab file path (if Kerberos authentication is used)

1. Log in to FineBI as the admin, choose System Management > Data Connection > Data Connection Management, and click New Data Connection, as shown in the following figure.

Note:

If you are not the admin, you can configure data connections only after the admin assigns you the permission on Data Connection Management. For details, see Data Connection Permissions Overview. 

2. Select the Apache Impala icon, as shown in the following figure.

3. Select Custom from the Driver drop-down list, select the driver uploaded in section "Version/Driver", and enter the connection information collected in section "Connection Information Collection." To use this data connection in the direct connection version, you need to suffix the URL with the parameter ;UseNativeQuery=1, as shown in the following figure.

For details about Kerberos authentication, see Kerberos Authentication in Data Connection.

If an Impala database requires authentication, you need to suffix the URL with the AuthMech parameter. Different parameter values represent different authentication methods, as shown in the following table.

Note:

If the database does not have a username or password, the URL does not need to be suffixed with this parameter.

The URL format is jdbc:impala://IP address:Port number/DB name;AuthMech=n. (n can be 0, 1, 2, or 3, representing the above-mentioned authentication methods respectively.)

Username and password authentication: If the value of the AuthMech parameter is 3, the URL format is jdbc:impala://IP address:Port number/DB name;AuthMech=3.

Username authentication: If the value of the AuthMech parameter is 2, the URL format is jdbc:impala://IP address:Port number/DB name;AuthMech=2.

Kerberos authentication: If the value of the AuthMech parameter is 1, the URL format is jdbc:impala://IP address:Port number/DB name;AuthMech=1.

If Kerberos authentication is used as the database authentication method, the URL needs to be suffixed with the AuthMech parameter, name of the client registered with the KDC, keytab file path, and URL corresponding to the Kerberos authentication (namely, AuthMech, KrbHostFQDN, and KrbServiceName parameters respectively).

The URL format is jdbc:impala://IP address:Port number/default;AuthMech=1;KrbHostFQDN=hostalias;KrbServiceName=impala.

For example: jdbc:impala://192.168.5.127:21050/default;AuthMech=1;KrbHostFQDN=quickstart.cloudera;KrbServiceName=impala. For details, see Kerberos Authentication in Data Connection.

4. Click Test Connection. If the connection is successful, click Save, as shown in the following figure.

Two methods are available to add database tables to FineBI.

• Adding Database Tables

• Adding SQL Datasets

Database Notice

• When you connect FineBI to Apache Impala, data query is case-insensitive. The query results returned are in lowercase.

• Impala does not support the date type, but supports the timestamp type.

• Impala does not support multiple distinct count statements. SQL errors may be reported when you preview data with multiple distinct count statements.

Connection Failure After Version Upgrade

Problem:

After the upgrade, the connection failed and the error message "Error initialized or created transport for authentication: java.io.IOException" is reported, as shown in the following figure.

Cause:

After the upgrade, the .ini file created by the built-in JASS code will be lost, causing a connection failure.

Solution:

Change the authentication method of Kerberos to the second one. For details, see Kerberos Authentication in Data Connection.