This chapter contains topics about installing and configuring the Cluster Server feature of Siebel Analytics.

NOTE:  The Cluster Server feature works for Siebel Analytics Server only. Analytics Web Server cannot be clustered in the same way as the Siebel Analytics Server, because metadata is stored in a single location (Web Catalog) on the Web server file system. However, see also the following Tip.

TIP:   Although the Analytics Web Server itself cannot be clustered, Siebel Analytics Web contains utilities to support replication of Web catalog content among multiple Analytics Web servers. For more information about these Web Replication utilities, see the Siebel Analytics Web Administration Guide.

For a list of all hardware requirements for clustered Siebel Analytics Servers, cluster controllers, and the shared file system used as the Siebel Analytics repository publishing directory, see Siebel System Requirements and Supported Platforms on Siebel SupportWeb.

Analytics Server Cluster Components

Figure 3 shows a typical small cluster server setup. A small cluster consists of the following components:

• Primary Cluster Controller. The role of the primary cluster controller is to monitor the operation of the servers in the cluster and to assign sessions within the cluster.
• Secondary Cluster Controller. The secondary cluster controller assumes the role of the primary cluster controller if the primary controller is unavailable.
• Master Server. The master server is a clustered Analytics Server to which the Administration Tool connects for online repository changes.
• Cluster Manager. The Cluster Manager is available in the Administration Tool when a repository is open in online mode. Server nodes can be stopped, started and inactivated from the Cluster Manager.
• Repository Publishing Directory. This directory is shared by all Analytics Servers participating in a cluster. It holds the master copies of repositories edited in online mode. The Clustered Analytics Servers examine this directory upon startup for any repository changes. The directory typically resides on a shared file system visible to all servers in the cluster.

  Figure 3. A Typical Cluster Controller Setup

Cluster Controller Functions

The cluster controller server performs the following functions:

• Services new ODBC requests:

  When a new ODBC request comes into either the primary or secondary cluster controller server, the following events occur:

  • One of the cluster controller servers responds with a Siebel Analytics Server name and port number.
  • The ODBC client uses this returned Siebel Analytics Server name and port number to connect to the Siebel Analytics Server directly.
  • The ODBC request goes through the Siebel Analytics Server CLIENT_SERVER_PORT (9703 in Figure 3).
  • The session assignment goes to the primary cluster controller's CLIENT_CONTROLLER_PORT (9706 in Figure 3).
• Monitors the health of the Siebel Analytics Server nodes that participate in this cluster:
  • Each Siebel Analytics Server node loads the repository, so the server knows which database to get data from.
  • Siebel Analytics Server nodes use a heartbeat mechanism over TCP/IP.
  • The Siebel Analytics Server nodes include a Server Manager.
  • The Server Manager starts, stops, and restarts the Siebel Analytics Server for the cluster controller server.