Lab 4.1: Planning a Server Cluster

Lab Objectives

After completing this lab, you’ll be able to

• Plan a server cluster
• Design resource groups and failover policies for the groups within that cluster

About This Lab

In this lab you’ll plan a server cluster for Wingtip Toys. You’ll be performing several of the steps described in Lesson 2 and Lesson 3: choosing a server cluster model, planning the resource groups for the cluster, and determining failover policies for the resource groups. For this lab, you can assume that some of the steps that are normally part of the cluster planning process have already been completed, as described in the scenario.

Before You Begin

Before you begin this lab, you must be able to

• Administer Windows 2000 Advanced Server
• Implement and administer the Cluster service in Windows 2000 Advanced Server

Scenario: Server Clusters for Wingtip Toys

Wingtip Toys is an online retail outlet that sells specialty toys over the Internet. The company, which markets toys to customers around the world, is a Web-based operation; they have no retail outlets or printed catalogs. As Wingtip Toys expanded into the international market, the company has seen a need to improve its Web site availability. The company has taken many steps to identify and reduce network risks by attempting to eliminate single points of failure. For example, the network now includes redundant components and network paths, multiple ISP connections, and UPS protection. Domain controller services have also been made highly available.

Wingtip Toys now wants to implement Windows clustering to support mail and database services. Exchange 2000 Server will be used for the mail services, and SQL Server 2000 will be used for the database services. The cluster will be made up of two Windows 2000 Advanced Server computers configured as member servers, and the cluster storage system will be made up of Fibre Channel connections and an offsite SAN to ensure data fault tolerance. In addition, the company wants to ensure that the services are highly available, but it wants to maximize the use of its hardware resources. The company isn’t overly concerned about performance degradation should failover occur because each server can adequately support all the services.

Exercise 1: Choosing a Server Cluster Model

In this exercise you must decide which model to use for your cluster. You can choose from a single-node configuration, active/passive configuration, or active/active configuration.

1. What are the differences between the three configuration models?
2. Which configuration best suits the needs of Wingtip Toys?

Exercise 2: Planning the Resource Groups

You should follow six steps to determine how to set up the resource groups for your cluster. In many cases the decisions to make in each step are very apparent because the scenario has clearly provided the information.

Listing the Server-Based Applications

In this step you should list the applications that will be running on the cluster.

1. What are those applications?

Sorting the List of Applications

Next, you should determine which applications can and will use failover.

1. Which applications will use failover?

Listing Other Resources

After you identify which applications will use failover, you should determine what other resources will be running on this cluster.

1. Which resources should you include?

Listing Dependencies

By now you should have a list of all of the resources that will be included in the resource groups. You must determine what dependencies exist among these resources.

1. How are these resources dependent on each other?

Making Preliminary Grouping Decisions

You should now group together the resources. At this point it should be clear how the resources will be grouped together.

1. How will resources be grouped together?
2. What are the advantages of this grouping strategy?

Making Final Grouping Assignments

After you’ve listed the resources that you want grouped together, you can assign a name to each group and create a dependency tree for each group.

1. How would you create a dependency tree for each group?

Exercise 3: Determining Failover Policies

Once you’ve planned your resource groups, you can decide how to configure the failover policies for each group. These policies will determine how the group behaves if failover occurs. You must configure three settings: Failover Timing, Preferred Node, and Failback Timing.

1. How should you configure the Failover Timing setting?
2. How should you configure the Preferred Node setting?
3. How should you configure the Failback Timing setting?