Installing the Elasticsearch server

The Elasticsearch component is installed in the INSTALL_DIR/elasticsearch-1.7.1 directory during the installation of IBM® Sterling™ Order Management System Software.

By default, the cluster name that is assigned to Elasticsearch is SSFS, but this name can be changed, if required.

Note: Elasticsearch requires at least Java version 7, but for Elasticsearch version 1.7.1 it is recommended that you use Oracle JDK version 1.8.0_25 or higher on the operating system that deploys the Elasticsearch server. Therefore, set JAVA_HOME accordingly.

You can deploy Elasticsearch on multiple servers, and multiple nodes of the Elasticsearch server can be started. To deploy Elasticsearch on a different server, copy the INSTALL_DIR/elasticsearch-1.7.1 directory to the desired server location. The new Elasticsearch node can then be started in the existing basic mode of configuration without making any changes.

Note: When deploying Elasticsearch on a new server instance, you must use the same cluster name that you use for other Elasticsearch server instances. Additionally, the new server instance must be located in the same network to ensure that the node is automatically discovered and can communicate with other nodes in the network.

The ES_HEAP_SIZE environment variable allows you to set the heap memory that will be allocated to the Elasticsearch Java process. It will allocate the same value to both min (-Xms) and max (-Xmx) values. Elasticsearch defaults these values to 256m and 1g, respectively. You need to set it to 4g.

Elasticsearch needs to keep lots of files open simultaneously, so you need to raise the maximum number of open files a user can have. To increase this setting, follow the instructions in the Elasticsearch documentation.

For information about Elasticsearch server properties and how to start and stop the Elasticsearch server instance, refer to the Elasticsearch documentation.