IBM SPSS Collaboration and Deployment Services 5.0: Clustering

Configure IBM SPSS C&DS 5.0 with an IBM WebSphere Cluster for High Availability and Scalability

The IBM® SPSS Collaboration and Deployment Services 5.0 product (also known as C&DS 5.0) is a JEE-based application that can be leveraged by other IBM SPSS products to provide enhanced capabilities. As a JEE-based application C&DS requires an application server to run on, and the logical choice is the IBM WebSphere® Application Server. This article describes how you can configure C&DS 5.0 to run on an IBM WebSphere Application Server Cluster in order to provide high availability and scalability.

Kevin Kepros, Senior Engineer, IBM

Kevin Kepros is a senior software engineer at the IBM Software Development Lab in Rochester, Minnesota. Kevin is a lead developer on the IBM SPSS Collaboration and Deployment Services product. Prior to joining the IBM Business Analytics team, Kevin was a lead developer on the WebSphere High Availability Manager component and a member of the WebSphere Clustering development team.



30 April 2013

Introduction

If you use C&DS 5.0, you know that it is a JEE-based application that requires an application server to function. But what do you do if your client load is extreme? What if your business depends on making the C&DS 5.0 application highly available? In this article, I describe how you can deploy C&DS 5.0 on an IBM WebSphere Network Deployment cluster to provide improved availability and scalability.

Clusters

For scalability, enterprise-quality JEE applications are usually deployed and executed in an IBM WebSphere Application Server cluster. The client requests are routed across the cluster, thereby distributing the work among all the application server processes that are members (see Figure 1):

Figure 1. Client requests distributed across a cluster
Client requests distributed across a cluster.

Fail over

In addition to scalability, deploying applications to an IBM WebSphere cluster can also provide high availability. If one cluster member fails, the router can direct client requests to the application on one of the remaining cluster members.


Overview of steps

Installing and configuring C&DS 5.0 5.0 to run on an IBM WebSphere Cluster involves completing the following steps:

  1. Create the desired IBM WebSphere topology (cell, cluster, servers, and so on).
  2. Install the C&DS 5.0 product.
  3. Configure C&DS 5.0 and indicate the cluster as your deploy target.
  4. For each node in the cluster, configure a shared directory to your IBM SPSS Collaboration and Deployment Services 5.0 installation.
  5. For each node in the cluster, verify or update the IBM WebSphere "CDS_HOME" location.

Let's take a closer look at each of these steps.


Step 1: Create IBM WebSphere cluster topology

This document is not intended to teach you about IBM WebSphere Application Server administration. I do review the typical administrative flow required to establish an IBM WebSphere cluster topology, however.

  1. Install IBM WebSphere Application Server (Network Deployment edition) on each of your desired nodes.

    TIP: For IBM SPSS Collaboration and Deployment Services usage, your IBM WebSphere Deployment Manager node and each node that will host a cluster member running the C&DS 5.0 application should be based on the same operating system type (for example, all Linux® nodes or all Microsoft® Windows® nodes).

  2. Select a node to host your IBM WebSphere administration console, and use the IBM WebSphere Profile Management Tool to create an IBM WebSphere Management Deployment manager profile on that node, as shown in Figure 2:
    Figure 2. IBM WebSphere Profile Management Tool: Deployment manager profile creation
    IBM WebSphere Profile Management Tool: Deployment manager profile creation

    TIP: In IBM WebSphere, a Deployment manager is the central administrative process for a cell. Do not confuse this with the C&DS 5.0 Deployment manager client application.

  3. For each node where you plan to host cluster members, use the IBM WebSphere Profile Management Tool to create a Custom profile and federate the node to the IBM WebSphere Deployment manager process, as shown in Figure 3:
    Figure 3. IBM WebSphere Profile Management Tool: Custom profile creation
    IBM WebSphere Profile Management Tool: Custom profile creation

    TIP: If you plan to federate your node during the Custom profile creation, you must have the IBM WebSphere Deployment manager running and provide the Deployment manager hostname and SOAP port information. Otherwise, you can federate the node after you create the Custom profile using the addNode command located in the profile bin directory (for example, <WAS_HOME>/profiles/Custom01/bin).

  4. After you have your nodes federated, you can use the IBM WebSphere administrative console (http://<WASDmgrHost>:9060/ibm/console) to create a cluster and members on your nodes, as shown in Figure 4 and Figure 5:
    Figure 4. IBM WebSphere Administrative console: Clusters panel
    IBM WebSphere Administrative console: Clusters panel
    Figure 5. IBM WebSphere Administrative console: New cluster pane
    IBM WebSphere Administrative console: New cluster pane

    For more information on IBM WebSphere administration tasks, see the link to the "WebSphere Application Server Version 8 - Information Center" in the resources section of this document.


Step 2: Install IBM SPSS Collaboration and Deployment Services 5.0

There are various prerequisites you must meet before completing the installation of the IBM SPSS Collaboration and Deployment Services 5.0 Server. They include the following:

  • Supported database product (for example, IBM DB2) installed and ready for use
  • Supported IBM WebSphere Application Server product installed and topology created
  • IBM Installation Manager 1.5 (or higher) installed
    • NOTE: If you are installing IBM SPSS Collaboration and Deployment Services from physical media (disc), then IBM Installation Manager will automatically be installed for you.

Launching the IBM SPSS Collaboration and Deployment Services Server installation

The initial launch of the C&DS 5.0 5.0 Server installation varies slightly depending on how you obtained the product.

TIP: When planning to deploy IBM SPSS Collaboration and Deployment Services on an IBM WebSphere Cluster, you must install the IBM SPSS Collaboration and Deployment Services product on the same system that hosts the IBM WebSphere Deployment Manager process.

Physical media

If you are installing from physical media (disc), you are presented with a LaunchPad screen, and you can select the option to install the IBM SPSS Collaboration and Deployment Services Server. The installation process is started and IBM Installation Manager is either installed (if not already installed), or the existing IBM Installation Manager install is used. In either case, the IBM SPSS Collaboration and Deployment Services 5.0 Server installation process is also started, as shown in Figure 6:

Figure 6. IBM SPSS Collaboration and Deployment Services 5.0 LaunchPad
IBM SPSS Collaboration and Deployment Services 5.0 LaunchPad

Downloaded content

If you downloaded the product files to a local directory, you need to install the IBM Installation Manager product and then configure a repository location in the IBM Installation Manager product to point at your downloaded content. This is done using the IBM Installation Manager -> File -> Preferences -> Repositories menu option.

For more information on installing IBM SPSS Collaboration and Deployment Services with IBM Installation Manager, see the link to the "Getting Started with IBM Installation Manager when Installing IBM SPSS Collaboration and Deployment Services Repository 5.0" in the resources section of this document.

IBM Installation Manager

Whether using physical media or downloaded files, the IBM Installation Manager product is used to complete the installation of the IBM SPSS Collaboration and Deployment Services Server. As with any typical software installation, you need to agree to the license, specify the desired installation directory, and so on. After the files have been installed to your system, you are presented with an option to launch the IBM SPSS Collaboration and Deployment Services Configuration Tool (see Figure 7):

Figure 7. IBM Installation Manager: Option to launch IBM SPSS Collaboration and Deployment Services 5.0 Configuration Tool
IBM Installation Manager: Option to launch IBM SPSS Collaboration and Deployment Services 5.0 Configuration Tool

At this point, the installation of the IBM SPSS Collaboration and Deployment Services 5.0 Server application is complete. You can begin the IBM SPSS Collaboration and Deployment Services 5.0 Server configuration process by having the C&DS 5.0 Configuration Tool launched automatically, or you can choose "None," exit the IBM Installation Manager, and manually launch the IBM SPSS Collaboration and Deployment Services Configuration Tool by running the <CDS_HOME>/bin/configTool command.


Step 3: Configure IBM SPSS Collaboration and Deployment Services 5.0 for a cluster

With a database available, the desired IBM WebSphere Application Server cluster topology created, and the IBM SPSS Collaboration and Deployment Services 5.0 product files installed, you are now ready to begin the IBM SPSS Collaboration and Deployment Services configuration step, shown in Figure 8:

Figure 8. IBM SPSS Collaboration and Deployment Services 5.0 Configuration Tool
IBM SPSS Collaboration and Deployment Services 5.0 Configuration Tool

TIP: Before attempting to configure IBM SPSS Collaboration and Deployment Services, be sure your database system is running and accessible, your IBM WebSphere cluster topology has been created, your IBM WebSphere administrative processes (Deployment Manager and Nodeagent(s)) are running, and your IBM WebSphere Application Server processes (Cluster and Members) are stopped.

IBM SPSS Collaboration and Deployment Services Configuration Tool

When configuring C&DS 5.0, the cluster-specific configuration is located on the Application Server panel of the C&DS 5.0 Configuration Tool. The other C&DS 5.0 Configuration Tool panels (Database, Existing data, Encryption, and so on) do not contain any cluster-specific information, so this article focuses on the Application Server panel.

Application Server panel

The first thing to notice about the Application Server configuration panel is that most of the entry fields are not editable. The entries on this panel are dynamic and change based on the selected application server type and topology, as shown in Figure 9:

Figure 9. Application Server configuration panel
Application Server configuration panel

IBM WebSphere profile directory

By default, IBM WebSphere is selected as your application server product. For our scenario, this choice is correct, so move ahead to the WebSphere profile directory field. Click Browse, and locate your IBM WebSphere Deployment Manager profile directory (for example, <WAS_HOME>/profiles/Dmgr01). When you save the profile directory information, the IBM SPSS Collaboration and Deployment Services Configuration Tool inspects the IBM WebSphere topology, as shown in Figure 10:

Figure 10. IBM WebSphere Application Server inspection indicator
IBM WebSphere Application Server inspection indicator

The remaining panel entries are dynamically updated based on the findings. For example, the profile information can automatically provide data on the IBM WebSphere installation directory, whether the topology is a single server or a managed cell, if IBM WebSphere administrative security is enabled, and more. See Figure 11:

Figure 11. IBM WebSphere Application Server panel: Dynamic update
IBM WebSphere Application Server panel: Dynamic update

Server topology

Depending on your topology, you have the option of selecting a managed server or a cluster. The choice varies slightly based on whether you have individual application servers, clusters, or both defined in your topology. For this scenario, select the cluster option if it is not already selected, as shown in Figure 12:

Figure 12. IBM WebSphere Application Server panel: Target server topology
IBM WebSphere Application Server panel: Target server topology

The name of the target IBM WebSphere cluster is automatically filled in if you only have a single cluster defined in the topology. If there are multiple clusters defined, then you need to use the drop-down list to select the desired target cluster.

URL prefix

The URL prefix is a C&DS 5.0 server configuration attribute that is indirectly related to the application server. For a C&DS 5.0 cluster deployment, it can be defined as "The URL where a server-initiated IBM SPSS Collaboration and Deployment Servicesweb service request should be routed". Typically, the cluster is fronted by some type of load balancer that is used to distribute requests across the cluster members. Using the URL prefix, server-initiated requests can be sent back to the load balancer for distribution across the cluster. Therefore, the URL prefix should be defined to "point at" your load balancer, for example:

http://<LB_host>:<LB_port>

TIP: If you don't have a load balancer configured, you can temporarily enter the information for one of the application servers defined in your C&DS 5.0 cluster. Specifying the hostname and HTTP port of one of your cluster members allows you to complete the C&DS 5.0 configuration. The C&DS 5.0 application functions, but any server-initiated requests are sent to the server that you provided. Later when you have a load balancer in place, you can update the value of the URL prefix using the browser-based IBM SPSS Collaboration and Deployment Services Deployment Manager client (for example:

http://<host>:<port>/security/login

Node

For a cluster, node-specific information is not required, and the entry field cannot be edited. If you were deploying to a managed single server and your topology contained multiple servers on multiple nodes, then you would need to provide both a server name and node name in order to uniquely identify your target application server. See Figure 13:

Figure 13. IBM WebSphere Application Server panel: Target node, administrative credentials
IBM WebSphere Application Server panel: Target node, administrative credentials

Server administrator

If the application server topology has administrative security enabled, the User name and Password fields would require the administrator credentials. In this example, I did not have application server administrative security enabled, so the fields cannot be edited.

Configuration summary

While it varies by topology, the minimum amount of application server information required to configure IBM SPSS Collaboration and Deployment Services for use in an IBM WebSphere cluster is the following:

  • IBM WebSphere Deployment Manager profile directory
  • URL prefix

If you have multiple clusters defined or IBM WebSphere administrative security enabled, you also need to provide:

  • Target cluster name
  • Administrator credentials (user ID, password)

For more information on the C&DS 5.0 Configuration Tool refer to the Repository Installation and Configuration Guide which is available at the "IBM SPSS Collaboration and Deployment Services v5.0 - Information Center" link provided in the Resources section of this document.


Step 4: Shared Directory

For a C&DS 5.0 cluster topology, the C&DS 5.0 product is installed on the IBM WebSphere Deployment Manager node. The application servers that are members of the cluster, however, are likely to be located on one (or more) other nodes. C&DS 5.0 requires that each node in the cluster have access to the original IBM SPSS Collaboration and Deployment Services product installation directory. Therefore, you must use the appropriate operating system commands (for example, mapped drive, mount point) to allow access to the original IBM SPSS Collaboration and Deployment Services product installation directory from each node. See Figure 14:

Figure 14. Clustered nodes and shared C&DS 5.0 installation directory
Clustered nodes and shared C&DS 5.0 installation directory

TIP: It is possible to eliminate the requirement for a shared directory by having a local copy of the original IBM SPSS Collaboration and Deployment Services product installation directory on each node. You, however, become responsible for keeping the replicated directories up-to-date. For example, if you install IBM SPSS Collaboration and Deployment Services service packs or IBM SPSS Collaboration and Deployment Services extensions that modify the original product installation directory, you need to replicate those changes to each node. Keeping the replicated copies in sync with the original is your responsibility.


Step 5: IBM WebSphere CDS_HOME

As part of the C&DS 5.0 Configuration process, a new IBM WebSphere environment variable named CDS_HOME was created. This variable is node scoped, which means that for every node in the target cluster, there is a CDS_HOME variable defined. After you have made the original C&DS 5.0 product installation directory available to each node, you should open the WebSphere administrative console and verify the setting of the CDS_HOME variable. The CDS_HOME variable for each node in the cluster should contain a valid path to the original C&DS product installation directory accessible from that node. Use the IBM WebSphere administrative console to verify or update the path information if necessary. See figures Figure 15, Figure 16, and Figure 17:

Figure 15. IBM WebSphere administrative console: Environment variables
IBM WebSphere administrative console: Environment variables
Figure 16. IBM WebSphere environment variables: CDS_HOME
IBM WebSphere environment variables: CDS_HOME
Figure 17. Node scoped CDS_HOME variable entry
Node scoped CDS_HOME variable entry

Conclusion

Deploying C&DS 5.0 to an IBM WebSphere Application Server cluster can provide improved availability and scalability. This article described the five steps that are required to configure this type of environment.

To summarize, follow these steps to deploy C&DS 5.0 on an IBM WebSphere cluster:

  1. Create the desired WebSphere topology (cell, cluster, servers, etc.).
  2. Install the C&DS 5.0 product.
  3. Configure C&DS 5.0, and indicate the cluster as your deployment target.
  4. For each node in the cluster, configure a shared directory to your IBM SPSS Collaboration and Deployment Services 5.0 installation.
  5. For each node in the cluster, verify or update the IBM WebSphere CDS_HOME location.

The installation and configuration of IBM SPSS Collaboration and Deployment Services 5.0 on an IBM WebSphere cluster is not much different from deployment to a single server. Armed with the knowledge provided in this article, you should be able to establish your own highly available and scalable C&DS 5.0 deployments.

Acknowledgments

I am grateful to the following people who contributed toward preparing or reviewing this article for its technical content and accuracy: Keith Swanson, IBM SPSS Collaboration and Deployment Services development, and Duane Wiebe, IBM SPSS Collaboration and Deployment Services quality assurance.

Resources

Learn

Get products and technologies

Comments

developerWorks: Sign in

Required fields are indicated with an asterisk (*).


Need an IBM ID?
Forgot your IBM ID?


Forgot your password?
Change your password

By clicking Submit, you agree to the developerWorks terms of use.

 


The first time you sign into developerWorks, a profile is created for you. Information in your profile (your name, country/region, and company name) is displayed to the public and will accompany any content you post, unless you opt to hide your company name. You may update your IBM account at any time.

All information submitted is secure.

Choose your display name



The first time you sign in to developerWorks, a profile is created for you, so you need to choose a display name. Your display name accompanies the content you post on developerWorks.

Please choose a display name between 3-31 characters. Your display name must be unique in the developerWorks community and should not be your email address for privacy reasons.

Required fields are indicated with an asterisk (*).

(Must be between 3 – 31 characters.)

By clicking Submit, you agree to the developerWorks terms of use.

 


All information submitted is secure.

Dig deeper into Big data and analytics on developerWorks


static.content.url=http://www.ibm.com/developerworks/js/artrating/
SITE_ID=1
Zone=Big data and analytics, Information Management
ArticleID=877433
ArticleTitle=IBM SPSS Collaboration and Deployment Services 5.0: Clustering
publish-date=04302013