Creating clusters using the job manager

In a flexible management environment, you can submit the Create cluster job. A cluster is a set of servers that you manage together to balance workload. The job creates a cluster that runs on a deployment manager node. The cluster becomes a managed resource of the job manager.

Before you begin

Start the job manager, the deployment manager, and the federated node to which you want to add a cluster.

If the deployment manager is not a managed node of the job manager, register the deployment manager with the job manager. Registering enables the job manager to manage the deployment manager and its federated nodes. To submit the Create cluster job, a deployment manager node must be a managed node of the job manager.

Determine how you want to configure the cluster:
  • Review topics on clusters and workload management, especially the information about setting cluster weights.
  • Determine whether you want a cluster to group application servers, proxy servers, or on-demand routers.
  • Determine the node to which to add the first cluster member. The target node must be a managed node of the job manager.
  • For a cluster of servers that spans multiple systems in a Sysplex and has stateful session beans with an activation policy of Transaction deployed in them, define the passivation directory on a hierarchical file system (HFS) that is shared across the multiple systems in the Sysplex.

Your ID at the job manager must be authorized for the administrator role or the operator role to submit jobs. When you submit a job, you can specify a user name and password for authentication and authorization at the target node or nodes. When you submit a job to multiple target nodes, the user name and password or the credentials for the submitter must apply to all the job targets.

About this task

A cluster enables you to manage a group of application servers as a single unit, and distribute client requests among the application servers that are members of the cluster. You might want to create a cluster if you need to:

You can use the administrative console of the job manager to create a cluster on one or more managed nodes. From the job manager console, choose the Create cluster job, specify job options, review the summary, and submit the job. This topic describes how to use the job manager console to submit the job.

Instead of using the job manager console, you can run the createCluster job script in the AdministrativeJobs command group. See the Administrative job types topic.

Procedure

  1. Click Jobs > Submit from the navigation tree of the job manager administrative console.
  2. Choose a job to create a cluster.
    1. Select the Create cluster job type from the list.
    2. Optionally specify a description of the job.
    3. Click Next.
  3. Choose the job targets on which to add the cluster.

    Select one or more deployment manager managed nodes to which you can add clusters.

    1. Select a group of nodes from the list, or select Node names.
    2. If you selected Node names, then specify a node name and click Add, or click Find and specify the chosen nodes on the Find nodes page.
    3. If user authentication is required, specify your user name and password.
    4. Click Next.
  4. Specify parameters for the create cluster job.
    1. For Cluster name, specify a cluster name that is unique on the target nodes.

      To see the names of existing clusters on the target nodes, click Jobs > Node resources . The Node resources page lists the clusters, servers, and applications on managed nodes. After you determine what cluster names are already used, return to the Specify job parameters page and specify a unique name for the cluster that you want to create.

      For example, suppose a managed deployment manager node, myNode01, has one cluster, cluster1. Specify a name that is unique to myNode01:
      cluster2
    2. For Prefer local, if you do not want to enable node-scoped routing optimization for the cluster, specify false. Otherwise, leave the field empty. The default value is true.

      When enabled, node-scoped routing optimization routes requests to the node on which the cluster resides.

    3. For Cluster type, specify the type of server cluster to create.

      If you want to create a cluster to group application servers, leave the field empty. The default type is APPLICATION_SERVER.

      Otherwise, specify PROXY_SERVER or ONDEMAND_ROUTER.

    4. If you are creating a cluster on a z/OS node, optionally specify a cluster short name.

      If you do not specify a value, the product generates a unique short name.

    5. Optionally, expand Additional job parameters and specify values that further define the cluster.
      Table 1. Additional Create cluster job parameters. Specify job parameters as needed.
      Parameter name Description
      Create domain To create a replication domain with a name set to the name of the new cluster, specify true. The product uses the replication domain for HTTP session data replication. The default value is false.
      Server node If you want to convert an existing server to the first member of the cluster, specify the name of a managed node that has the existing server.
      Server name If you want to convert an existing server to the first member of the cluster, specify the name of the existing server that resides on the node specified for Server node.
      Member weight If you want the new cluster member to have a weight value other than 2, the default, specify a different weight value. A valid value is a number between 0 and 100. The weight controls the amount of work directed to a server. If the weight is greater than the weight assigned to other cluster members, the server receives a larger share of the workload.
      Node group If you want all cluster member to belong to a node group, specify the name of the node group.
      Replication entry If Create domain is set to true, optionally specify true to create a replicator entry for this member in the cluster replication domain. The default value is false.
    6. Click Next.
  5. Schedule the job.

    The times and dates that you specify are relative to the job manager.

    1. Optionally specify one or more e-mail addresses where notifications are sent when the job finishes.

      If you specify multiple e-mail addresses, separate them with commas.

    2. Select when the job is available for submission.
    3. Select the job expiration.
    4. Optionally specify a recurring interval for the job, a start date and time for the interval, and an end date and time for the interval.
    5. Click Next.
  6. Review the summary, and click Finish to submit the job.

Results

The target nodes run the job and attempt to create a cluster.

What to do next

On the Job status page, click the ID of the create cluster job and view the job status. Click the status refresh icon Specifies the refresh view icon for job status to refresh the displayed status.

If the job is not successful, view any error messages that result from running the job, correct the error condition, and submit the job again.

If the job is successful, click Jobs > Node resources to see the new cluster in the list of resources. The status of the cluster is Stopped.

After creating the cluster, you can run a job that creates a cluster member. You cannot start the cluster unless it has at least one cluster member.




In this information ...


IBM Redbooks, demos, education, and more

(Index)

Use IBM Suggests to retrieve related content from ibm.com and beyond, identified for your convenience.

This feature requires Internet access.

Task topic    

Terms of Use | Feedback

Last updated: Oct 22, 2010 12:21:29 AM CDT
http://www14.software.ibm.com/webapp/wsbroker/redirect?version=compass&product=was-nd-zos&topic=tagt_jobmgr_create_cluster
File name: tagt_jobmgr_create_cluster.html