Acceldata
 ODP
Search documentation… ⌘K

Access and Use Kudu Tables

You can access and interact with Kudu tables using several methods.

Kudu CLI Tool

The Kudu command-line tool is available at:

  • /usr/odp/$(odp-select --version)/kudu/bin/kudu
  • /usr/bin/kudu

To list tables in your cluster, run:

kudu table list $COMMA_SEPARATED_LIST_OF_MASTERS

Impala

Kudu integrates with Impala for querying and managing tables.

  • When using the Impala Mpack, ensure the checkbox impala_disable_kudu is unchecked to enable Kudu integration.

  • To create Kudu tables in Impala, specify the master addresses in the table properties as a comma-separated list.
  • In ODP version 3.3.6.1-1 or newer, this step is optional; Impala automatically sets the master addresses.

Example:

CREATE TABLE your_kudu_table (
id INT PRIMARY KEY,
name STRING
)
PARTITION BY HASH(id) PARTITIONS 3
STORED AS KUDU
TBLPROPERTIES (
'kudu.master_addresses' = '<FQDN_master_1>:<port>,<FQDN_master_2>:<port>,<FQDN-master_3>:<port>',
'kudu.num_tablet_replicas' = '3'
);
For details on using Kudu with Impala, see the Kudu documentation.

Kudu APIs

  • Kudu provides C++, Java, and Python client APIs. For details, see Developing Applications With Apache Kudu.
  • When installing Kudu using the Mpack, Python libraries will be installed for the version of Python supported by your ODP version.

Hive

  • Partial storage handler support is available.
  • Note: There may be issues if Kerberos authentication is enabled in Kudu.