HP P6000 Continuous Access Implementation Guide (T3680-96431, August 2012)

• Cannot be a snapshot
• Cannot be a mirrorclone
• Must be in normal operational state
• Must use mirrored write cache (the default)
The maximum number of virtual disks in a DR group and the maximum number of DR groups per
array vary with controller software versions. For current supported limits, see the HP P6000 Enterprise
Virtual Array Compatibility Reference.
Adding members to a DR group
When adding new members to an existing DR group, it is important to ensure that I/O consistency
of the destination virtual disks in the DR group is maintained at all times. This ensures data
consistency should it become necessary to fail over to the destination array.
To ensure I/O consistency of the destination array, the addition of new DR group members and
the accompanying normalization should occur before the new members are presented to an
application server for use. If the new virtual disks are used by an application server before they
are added to the DR group, when normalization begins the destination DR group members will be
in a data-inconsistent state until the new members have completed normalization. By adding the
new members to the DR group and allowing normalization to complete before presenting them to
the application server, I/O consistency is maintained at all times on the destination volumes in the
DR group.
Use the following steps when adding virtual disks to a DR group to ensure that the data on the
destination array remains I/O consistent at all times. These steps apply to all synchronous and
enhanced or basic asynchronous DR groups using XCS versions 6.200 or later.
1. Add the new virtual disks to the DR group.
The normalization process begins and should be allowed to complete before presenting the
new members to the application server for use. During normalization, the virtual disk members
on the destination array will be I/O consistent to the application. Because the application
server has not yet had the new members presented to it, it cannot write to the new members.
Consequently, the application server does not need the new members for an I/O consistent
view of application data if a failover to the destination array is required.
2. When normalization of the new members is complete, present the virtual disks to the application
server.
Data now written to the new virtual disks will be I/O consistent within the DR group because
of the guaranteed write ordering for multiple member DR groups.
Selecting replication mode
Select the replication mode when you create a DR group. Synchronous mode is the default, providing
identical copies on local and remote arrays as long as live replication is occurring. Asynchronous
mode may provide faster response to host server requests, but it does not ensure that data has
been replicated to the remote array before the host is told that the I/O is complete. The choice of
write mode has design implications and depends on business requirements. For detailed descriptions
of synchronous and asynchronous modes, see “Choosing a write mode” (page 20).
Specifying DR group write history log location and size
You can specify the size and location of the DR group write history log when you create a DR
group. The log size and location are dependent on the controller software version. For more
information on DR group write history log, see “Planning for DR group write history logs” (page 42).
Creating DR groups 99