Failover Clustering
Applies To: Windows Server 2008
A failover cluster is a group of independent computers that work together to increase the availability of applications and services. If one of the clustered servers fails, another server begins to provide service (a process known as failover).
Hierarchy of Managed Entities
.gif)
Managed Entities
| Name | Description |
|---|---|
A failover cluster is a group of independent computers that work together to increase the availability of applications and services. If one of the clustered servers fails, another server begins to provide service (a process known as failover). |
|
A failover cluster node is a server that is a member of a failover cluster. When a node is online, it is an active member of the cluster and can own and run clustered services and applications, honor cluster database updates, contribute votes to the quorum algorithm, and maintain heartbeats. |
|
In a failover cluster, the network adapter or adapters (also know as network interface cards or NICs) on each node must function correctly, as must other related network devices. Network connectivity is needed for the nodes to maintain communication with each other and with clients. |
|
In a failover cluster, the network or networks must function correctly so that the nodes can maintain communication with each other and with clients. |
|
A service or application configured within a failover cluster consists of multiple clustered resources, and uses configuration information that tells the cluster the order in which to start or stop the resources. The successful function of the clustered service or application depends on whether necessary clustered resources can come online. |
|
A clustered file server instance is a configuration through which you can share folders and files that reside on one or more disks in the cluster storage. To clients, the name of the server through which the folders and files are shared always appears the same. However, within the cluster, any node can own the file server instance at a given time. |
|
A clustered WINS server instance or DHCP server instance can be hosted by any node in the cluster at a given time, which can increase the availability of WINS or DHCP. |
|
A clustered DFS namespace instance can be hosted by any node in the cluster at a given time, which can increase the availability of the DFS namespace. |
|
A Generic Script Instance uses a script that monitors and controls a server application. The correct functioning of the script is very important in determining whether the combination of Cluster service, script, and application will work together to increase availability. |
|
A clustered Generic Service instance can increase the availability of a service that is not "cluster-aware" (not originally designed to run in a cluster). |
|
A clustered Generic Application instance can increase the availability of an application that is not "cluster-aware" (not originally designed to run in a cluster). |
|
Failover clusters with an even number of nodes usually use a quorum configuration that includes a "witness disk" or "witness file share." In a cluster with an even number of nodes and a quorum configuration that includes a witness, when the witness remains online, the cluster can continue sustain failures of half the nodes. If the witness goes offline, the same cluster can sustain failures of half the nodes minus one. |
|
In a failover cluster, the storage must function correctly so that each node can read and write cluster configuration information or data. Only one server at a time should be able to access a given disk (volume) in the cluster storage. Configuration or hardware problems that prevent access to the storage can disrupt the function of the cluster. |