Partilhar via


Server Cluster groups

Applies To: Windows Server 2003, Windows Server 2003 R2, Windows Server 2003 with SP1, Windows Server 2003 with SP2

Groups

A group is a collection of cluster resources with the following characteristics:

  • Groups define the units of failover. That is, when one resource in a group fails and it is necessary to move the resource to an alternate node, all of the resources in the group are moved to the alternate node. For more information, see Failover and failback.

  • A group is always owned by one node at any point in time. Likewise, a resource is always owned by a single group. These relationships ensure that all of a group's members reside on the same node.

For more information on resources or nodes, see Server Cluster Resources or Nodes.

Groups enable resources to be combined into larger logical units. Typically, a group is made up of related or dependent resources, such as applications and their associated peripherals and data. Indeed, groups define dependency boundaries for resources: resources cannot depend on resources in other groups. Because most application-level resources have some dependency on storage, most groups are configured around storage-class resources. A simple rule of thumb is this: If a resource uses shared storage resources, the resource belongs in the same group as that storage resource.

However, groups can also be established with resources that are unrelated and nondependent to balance the load or for administrative convenience. For more information on dependent resources, see Resource dependencies. For more information on setting resource dependencies, see Modify resource dependencies.

Every group maintains a prioritized list of the nodes that can act as its host. The preferred nodes list is generated from two sources:

  • The Cluster service produces a list of preferred nodes for a group from the list of possible owners that is maintained by the group's resources. For more information on Cluster service, see Cluster service.

  • Administrators can add nodes to the list through a cluster-management application, such as Cluster Administrator. For more information on using Cluster Administrator to specify the preferred owners of a group, see Specify preferred owners of a group.

To fully exploit the processing power of a cluster, establish at least as many groups as there are nodes in the cluster and evenly distribute these groups across the cluster.

The following topics discuss important concepts related to groups: