Hewlett-Packard White Paper Sample

The primary purpose of a failover cluster (also called server cluster) is to increase application availability, but at the same time it ensures that only one instance of an application is online at any point-in-time. The mechanism of ensuring that there are never two instances of one application running is one of the core technologies of the server cluster. The Microsoft Failover Clustering implementation is the quorum mechanism. It also supports geographically dispersed clusters. Microsoft Failover Clustering supports shared access quorum disk and the majority node set quorum approach.

HP StorageWorks Enterprise Virtual Array Cluster Extension (EVA CLX) for Windows extends a Microsoft cluster across two sites. This improves availability by providing disaster tolerance in case of core component failures such as storage systems and site disasters like fire, power outage, and even larger natural disasters. EVA CLX supports only majority node set quorum clusters.

In the clustered CLX environment, users can have point-in-time copies of the application disks using the EVA Business Copy feature (either mirror clone, snap clone, or snapshots) in the local EVA array where the application disks are replicated using the EVA Continuous Access (CA). If there is a data inconsistency in the application source disks, it can be recovered by doing the instant restore using the point-in-time copies. In this scenario, this document explains the considerations required to be followed in HP StorageWorks EVA Cluster Extension configurations, before and after taking the point-in-time copies and instant restore from the point-in-time copies.