The following sample configuration represents a campus cluster with two sites, Site A and Site B.

With SFW, a campus cluster can be set up using a Cluster Server (VCS) configuration. Both configurations involve setting up a single cluster with two nodes that are in separate buildings and are connected via a single subnet and Fibre Channel SAN. Each node has its own storage array with an equal number of disks and contains mirrored data of the storage on the other array. SFW provides the mirrored storage and the disk groups that make it possible to fail over the storage by deporting the disk groups on one node and importing them on the other.

If a site failure occurs in a two-node campus cluster, the remaining cluster node will not be able to bring the cluster disk groups online because it cannot reserve a majority of disks in the disk groups. To allow for failover to the other site, a procedure forces the import to the other node, allowing a cluster disk group to be brought online on another node when that node has a minority of the cluster disks.

Implementing these force import procedures should be done with care. The primary site may appear to have failed but what really has happened is that both the storage interconnect between sites and the heartbeats have been lost. In that case, cluster disk groups can still be online on the primary node. If a force import is done so that the data can be accessed on the secondary site, the cluster disks will be online on both sites, risking data corruption.