Implementing Analytic Engine Clustering

The following procedure describes the high level steps required to configure Optimize for Analytic Engine clustering with links to the appropriate sections of this chapter containing more specific information. Before completing these steps, you should understand your goals for clustering and make certain that you have the appropriate hardware available and that it is configured appropriately. Also, you must have a Terracotta Server Array set up and configured. Refer to the Terracotta documentation and web site for information about setting up and configuring a TSA.

When you first implement clustering, or if you change from a clustered to a non-clustered system, you must configure your Analytic Engines and deploy your configuration as described in the following procedure. Note that you can add or remove Analytic Engines in an existing cluster, without redeploying your environment.

Note:
Each Analytic Engine in a cluster must have a unique identity; that is to say, each Analytic Engine must have its own logical server. When you create a logical server, the system creates a unique ID and that ID is pushed to the Analytic Engine when the environment is deployed. If you configure multiple Analytic Engines that share a single ID, they will not function in a clustered environment.

To implement an Analytic Engine cluster:

1. On the Central Configuration Design Servers tab, add the appropriate number of Analytic Engine logical servers for your system. For more information, see Adding Logical Servers to an Environment.

2. Review Analytic Engine subcomponent settings for each logical server.

Ensure that all (global configuration) settings across the Analytic Engine logical servers are the same. For instance, if you decide to change the "Data Maintenance Settings", you must make the identical change for each Analytic Engine logical server in the cluster.

Also, ensure that the TSA URL is populated in either the default TSA configuration or each Analytic Engine logical server TSA configuration. For more information, see Defining Analytic Engine Cluster Settings.

3. Define the appropriate hardware hosts for your Analytic Engine cluster as appropriate on the Define Hosts tab. For more information, see Defining Host Servers for an Environment.

4. Map your Analytic Engine logical servers to the desired host server machines as appropriate on the Map Servers tab. For more information, see Mapping Logical Servers.

5. Map endpoints as appropriate for your system configuration. For more information, see Mapping Endpoints. When mapping endpoints for a cluster you should consider the following:

When deploying more than one Analytic Engine to the same host, ensure that each Analytic Engine uses unique ports for the Configuration Agent and WS Registry under the "INCOMING CONNECTIONS".

When deploying multiple Analytic Engines to unique hosts, each Analytic Engine can use the same port numbers.

The JMS provider must be the same across all Analytic Engines participating in the cluster.

6. Map the database pools as appropriate for your cluster. For more information, see Mapping Database Pools.

7. Validate your configuration as described in Validating an Environment Configuration.

8. Deploy your configuration to the host systems as described in Deploying an Environment.