Configuring the nodes in a cluster

The first time that you configure a cluster, you must use the startCluster command with the database update value set to true.

As part of a default cluster configuration, certain values in the database for service or adapter configurations, default document storage type, etc., are updated to get the cluster working. The default settings include no shared or mounted file system available with “line of sight” from all cluster nodes, etc. Certain service or adapter configurations are forcibly deployed on node1 and default document storage type is set up to “Database” for all business processes.

After you install the cluster and evaluate the customer configurations and requirements, the above conditions might change and custom configurations will be incorporated. To keep these custom configuration changes from being overwritten, the following cluster configuration script has an option to update the database:

The startCluster.sh nodeNumber command on node 1 automatically updates the database. The startCluster.sh nodeNumber command on all other nodes does not update the database.

You must use startCluster command with the true option only during the initial configuration of the cluster. If you must use the startCluster command again, use it with the false option. The false option prevents any configuration changes from affecting the system, especially after installation of a Fix Pack.

Important: You must run the startCluster command after installing the first node (node 1) on the host where you have installed the node. For v5.2.6 or later versions, after you run the startCluster command for the first node, the subsequent nodes are automatically clustered by the installer when they are installed. You must not execute the startCluster command on nodes other than node 1.

You must complete the following in a cluster setup, after installing the first node (node 1) on the host where node 1 is installed:

Windows operating system:

  1. Go to \<install_dir>\bin directory.

  2. Enter startCluster.cmd 1 true.

Linux/Unix operating system:

  1. Go to /<install_dir>/bin directory.

  2. Enter ./ startCluster.sh 1 true.

Note: You must install the subsequent nodes 1 at a time and ensure the node installation is complete and successful before continuing with the next node.