2. Installing VectorH : Installation Overview
Share this page                  
Installation Overview
You install VectorH on one DataNode of the Hadoop cluster. This node becomes the VectorH master node. The master node drives the setup of the VectorH slave nodes. The installation process installs VectorH, the Intel MPI Runtime, a Java Runtime Environment (JRE), and sets up the HDFS root and data locations.
Best Practice — We recommend installing VectorH on at least three DataNodes (one of which is the VectorH master node).
By default, VectorH is installed on all DataNodes of the cluster. If you want to install it on a subset of DataNodes, you must use a two-step installation process:
1. Perform installation on the master node only.
2. Set up the slave nodes manually (see Set Up Slave Nodes Manually).
Step 2 can be repeated at any time to reconfigure which nodes are used as slaves.