Sqoop 2 is distributed as two separate packages: a client package (sqoop2-client) and a server package (sqoop2-server). Install the server package on one node in the cluster; because the Sqoop 2 server acts as a MapReduce client this node must have Hadoop installed and configured.

Install the client package on each node that will act as a client. A Sqoop 2 client will always connect to the Sqoop 2 server to perform any actions, so Hadoop does not need to be installed on the client nodes.

Depending on what you are planning to install, choose the appropriate package and install it using your preferred package manager application.

  Note: The Sqoop 2 packages can't be installed on the same machines as Sqoop1 packages. However you can use both versions in the same Hadoop cluster by installing Sqoop1 and Sqoop 2 on different nodes.

To install the Sqoop 2 server package on a Red-Hat-compatible system:

  Important:

If you have not already done so, install Cloudera's yum, zypper/YaST or apt repository before using the following commands to install Sqoop 2. For instructions, see CDH 5 Installation.

$ sudo yum install sqoop2-server

To install the Sqoop 2 client package on an Red-Hat-compatible system:

$ sudo yum install sqoop2-client

To install the Sqoop 2 server package on a SLES system:

$ sudo zypper install sqoop2-server

To install the Sqoop 2 client package on an SLES system:

$ sudo zypper install sqoop2-client

To install the Sqoop 2 server package on an Ubuntu or Debian system:

$ sudo apt-get install sqoop2-server

To install the Sqoop 2 client package on an Ubuntu or Debian system:

$ sudo apt-get install sqoop2-client

  Note:

Installing the sqoop2-server package creates a sqoop-server service configured to start Sqoop 2 at system startup time.

You are now ready to configure Sqoop 2. See the next section.