HP EVA Cluster Extension Software Administrator Guide (5697-2445, July 2013)

Figure 1 Physical replication using HP EVA Continuous Access
EVA Continuous Access-mirrored disks have a read/write-enabled source (local) disk and a read-only
destination (remote) disk. Current cluster software products cannot distinguish between read-only
and write-enabled disks, and cannot enable disk access if the disk is not write-enabled during the
server boot process.
With EVA Cluster Extension, the consistency and concurrency of the data can be checked when
the resource comes online in the remote data center. The capability for restoring the application
service after a failure of the server is called disaster tolerance.
Automated redirection of mirrored disks
Storage systems with EVA Continuous Access automatically redirect the mirroring destination. This
means that EVA Continuous Access almost instantaneously swaps the source/destination relationship
of DR group members if the application must access the destination disk (vdisk). This feature ensures
that the disks are always accessible when failover to a remote data center occurs.
NOTE: Vdisks are virtual disks used in a HP EVA storage system for the storage of application
data. A DR group includes one or more vdisks in an EVA storage system. The DR group is the unit
in which EVA Cluster Extension failover/failback operations are performed. For more information
about DR groups, see the HP EVA Continuous Access Implementation Guide.
If the links between your storage systems are broken, depending on the configuration, the source
EVA array can log all the write I/Os of a DR group member to a write history log and replicate
the changes later after the links become available. In case of a failover, EVA Cluster Extension
takes the appropriate action for each link/array status and makes sure that your application service
has the latest data.
Cluster setup considerations
For quorum considerations, see “MSCS” (page 8).
Automated redirection of mirrored disks 7