Starting clusters

You can start all members of a cluster at the same time by requesting that the state of a cluster change to running. That is, you can start all application servers in a server cluster at the same time.

Before you begin

Make sure that the members of your cluster have the debug port properly set. If multiple servers on the same node have the same debug port set, the cluster could fail to start. Read about how to change the debug port in the topic about Java™ virtual machine settings.

[Windows] If you use the Windows Services facility to start and stop application servers that are part of a cluster, remember that the cluster state does not always update correctly. For example, if a cluster is running and you stop a cluster member through the Services GUI, the cluster state remains as started even though the server is no longer running.

If you want cluster member components to dynamically start as they are needed by the installed applications, verify that the Start components as needed option is selected in the configuration settings for each of the cluster members before you start the cluster. Selecting this option can improve cluster startup time, and reduce the memory footprint of the cluster members. Starting components as they are needed is most effective if all of the applications that are deployed on the cluster are of the same type. For example, using this option works better if all of your applications are Web applications that use servlets, and JavaServer Pages (JSP). This option works less effectively if your applications use servlets, JSPs and Enterprise JavaBeans (EJB).

Avoid trouble: To ensure compatibility with other WebSphere® products, the default setting for this option is deselected. Before selecting this option, verify that any other WebSphere products, that you are running in conjunction with this product, support this functionality.

About this task

When you request that all members of a cluster start, the cluster state changes to partially started and each server that is a member of that cluster launches, if it is not already running. After all members of the cluster are running, the cluster state becomes running.

[z/OS]Avoid trouble: From the z/OS® MVS™ console, you must individually start each server that you want to run. With the administrative console, you can start each server individually, or you can start a cluster. Starting a cluster automatically starts all of the servers that are defined as part of that cluster.

Procedure

  1. In the administrative console, click Servers > Clusters > WebSphere application server clusters > .
  2. Select the clusters whose members you want started.
  3. Click Start or Ripplestart.
    • Start launches the server process of each member of the cluster by calling the node agent for each server to start the servers. After all servers are running, the state of the cluster changes to running. If the call to a node agent for a server fails, the server does not start.
    • Ripplestart combines stopping and starting operations. It first stops and then restarts each member of the cluster. For example, your cluster contains 3 cluster members named server_1, server_2 and server_3. When you click Ripplestart, server_1 stops and restarts, then server_2 stops and restarts, and finally server_3 stops and restarts. Use the Ripplestart option instead of manually stopping and then starting all of the application servers in the cluster.
    Avoid trouble: If recently added cluster members do not start, you might not have selected Synchronize changes with nodes when you added the members to the cluster. To determine if this is the problem:
    1. In the administrative console click Servers > Clusters > WebSphere application server clusters > , select the cluster whose members did not start, and click Stop.
    2. Click the name of the cluster, click OK, and then click Review.
    3. Select Synchronize changes with Nodes, and then click Save.
    4. Start the cluster and verify that all of the cluster members now start.

Results

When you start the members of a cluster, you automatically enable workload management.