White Papers

13 Dell Storage PS Series Arrays: Scalability and Growth in Virtual Environments | TR1072
8 Replication and scale
It is important to understand how replication plays into the connection-count limitations and design of a DR
site. PS Series firmware allows replication reserve space to be selected from a single pool at the destination
site. Recovery operations must be planned accordingly because any or all volume(s) from any pool in Site A
can be replicated to Site B, but there must be enough replication reserve in the pool designated at Site B to
accommodate replication activities from Site A.
Note: If replication space capacity is of concern, PS6610 series arrays are recommended due to their high
density storage capacity.
When planning full site replication and recovery, thought must be used in planning the recovery strategy at
Site B. With the possibility of replicating volumes from up to 4 concurrent pools from Site A, careful attention
must be taken to understand the number of connections that may be needed during a recovery operation at
Site B and create the storage environment accordingly.
One approach is to add up the total number of connections to the PS Series SAN at Site A. If the total
connection count is within the limits of a single storage pool, nothing else is needed. Be mindful of any
additional volumes that may be created in the future and added to the replication plan. If the total connections
are above the maximum number supported for a single pool, then the best approach for Site B is to create
multiple PS Series Groups to ensure enough connections for each volume to recover.
Another approach is to mimic each pool at Site A with a separate PS Series Group at Site B. What this means
is that the PS Series SAN at Site A has pools 1–4. At Site B, there would be 4 separate PS Series Groups
created. The design would look similar to: Site A/Pool 1 would replicate to Site B/Group 1 DR, and Site A/Pool
2 to Site B/Group 2 DR and so on. This will allow a similar mirrored environment to be established for DR.
If an active/active (bi-directional replication) environment is configured or if one site has fewer resources than
the other, use careful consideration when designing recovery plans to make sure that on a failover, the
environment can handle the connection load required.
PS Series arrays, ASM/ME, ASM/VE, and VMware Site Recovery Manager support bi-directional replication
between PS Series SANs which means Site A can have virtual machines that are replicated to Site B and Site
B can have virtual machines that are replicated to Site A. A disaster at either site will allow for the VMs at the
other site to be brought online quickly, but this requires both sites to have sufficient resources to not only host
their current environment but the recovery environment as well. This can be factored in by determining which
VMs require higher priority as part of the recovery planning step when sizing the two sides.