次の方法で共有


Design for a Failover Cluster in Which All Nodes Run Hyper-V

Applies To: Windows Server 2008

If you want to consolidate multiple servers (as virtual machines) on one physical server, but want to avoid causing that server to become a single point of failure, you can create a failover cluster in which all servers (nodes) run Hyper-V and are configured to run one or more virtual machines as needed.

For designs that do not involve Hyper-V, see Mapping Your Deployment Goals to a Failover Cluster Design.

Important

  • If you are creating a failover cluster design in which the nodes run Hyper-V, be sure to review Requirements and Recommendations for Failover Clusters in Which All Nodes Run Hyper-V.

  • When you create a failover cluster where virtual machines running on one physical server can fail over to another physical server as needed, you are clustering the physical servers and virtual machines, not the applications that are running in a given virtual machine. The failure of a physical server (a cluster node) would cause another physical server to take over support of the virtual machines, but the failure of an application within a virtual machine would not. For increased availability beyond what the failover cluster provides, implement close monitoring of the applications as well, so that you can provide a quick response if an application fails.

  • Note that for the maximum availability of any server, it is important to follow best practices for server management—for example, carefully managing the physical environment of the servers, testing software changes before placing them into production, and carefully keeping track of software updates and configuration changes on all clustered servers.

    A failover cluster usually includes a storage unit that is physically connected to all the servers in the cluster, although any given volume in the storage is only accessed by one server at a time. The following diagram shows a simple two-node failover cluster that is connected to a storage unit, where the cluster consists of servers running Hyper-V and providing several virtual machines to clients.

    Failover on a two-node cluster where both nodes run Hyper-V

    To learn more about this design, see Example, Failover Cluster in Which All Nodes Run Hyper-V.

    For information about configuring this design, see Checklist: Failover Cluster in Which All Nodes Run Hyper-V (https://go.microsoft.com/fwlink/?LinkId=129123).