V11 SP8

Creating a User-Defined Instance for Hadoop

While creating a new pseudo-client, you also create a user-defined instance.

You can create additional instances to manage different Hadoop clusters.


  1. From the CommCell Browser, expand Client Computers > pseudo-client > Distributed Apps.
  2. Right-click Distributed Apps, point to All Tasks, and then click Create New Instance.

    The Create New Hadoop Instance dialog box appears.

  3. Enter the HDFS URI host and user details to connect to the Hadoop cluster:
    1. In the Instance Name box, type the instance name.
    2. In the HDFS URI box, type the Fully Qualified Domain Name (FQDN) of the Hadoop NameNode.

      Important: If you select the default HDFS URI, the URI details are fetched from the core-site.xml file. The core-site.xml file is located under the hadoop_installation_directory/conf/ directory.

    3. Optional: In the HDFS User box, type the Hadoop user name if a non-root user account is used to manage the Hadoop cluster.
    4. In the Hadoop Native Library Path box, type or click Browse to specify the path to the Hadoop native library (libhdfs.so).





    5. In the JVM Library Path box, type or click Browse to specify the path to the JVM library (libjvm.so).





  4. On the Hadoop tab, specify the data access nodes to add to the instance:
    1. In the Master Node list, select one of the data access nodes as a master node for the instance.
    2. Under Data Access Nodes, select the data access nodes that you want to add to the instance, and then click Add.
    3. In the Number of Data Readers box, enter the number of data streams.

      Tip: For optimal sharing of the backup load, the number of data readers must be greater than the number of data access nodes.

  5. On the Storage Device tab, select a storage policy from the Storage Policy list.
  6. Optional: To create a new storage policy, click Create Storage Policy, and then follow the instructions in the storage policy creation wizard.
  7. Optional: To perform LAN-free backups and restores, select a grid storage policy.

    For more information, see GridStor® (Alternate Data Paths) - Overview.

  8. Click OK.

A new instance is created with the name that you specify. A default subclient is created for the new instance.