When planning replication links between storage devices, comply the following rules:
l
Ensure that each controller of the primary storage device has a replication link to the
secondary storage device, which reduces I/Os forwarded between controllers. Otherwise,
ensure that each controller enclosure of the primary storage device has a replication link
to the secondary storage device.
l
A maximum of eight available links can be identified between a controller and the
remote device. When you add a remote device, a maximum of eight links can be used for
replication.
Data Planning
l
Capacity planning
–
Storage pool: Before configuring the remote replication feature, reserve 5
‰
of the
LUN capacity in the storage pool where the primary and secondary LUNs reside. If
you want to configure multiple value-added features on the primary or secondary
LUN, reserve 1% of the LUN capacity in the storage pool. In addition, for
asynchronous remote replication, reserve certain capacity in the storage pool to
which the primary and secondary LUNs belong for pre-write copy. For details about
space planning, see
9.7 Functions of Reserved Space for an Asynchronous
–
LUN: The primary LUN and secondary LUN must have the same capacity.
l
Bandwidth planning
–
For synchronous remote replication, a write success response is returned only after
the data in each write request is written to the primary site and secondary site. If the
primary site is far away from the secondary site, the write latency of applications is
quite long, affecting user services. Therefore, synchronous remote replication
applies to scenarios where the primary site is near to the secondary site, for
example, intra-city disaster recovery. Normally, the disaster recovery distance of a
metropolitan area network (MAN) is smaller than 200 km and the minimum
network bandwidth must be or larger than 64 Mbit/s. The transmission latency must
be smaller than 1 ms (single-way) and the actual network bandwidth must be larger
than the peak service I/O write bandwidth.
–
For asynchronous remote replication, the write latency of applications is
independent of the distance between the primary and secondary sites. As a result,
asynchronous remote replication is applied in the disaster recovery scenario where
the primary and secondary sites are far away from each other, or the network
bandwidth is limited. Generally, there is no explicit limit on the WAN distance
between the primary and secondary sites. It is recommended that the link bandwidth
be greater than 10 Mbit/s (peak bandwidth/bandwidth usage), and the unidirectional
transmission latency be smaller than 50 ms.
OceanStor V3 Series
Remote Replication Feature Guide for Block
2 Planning
Issue 06 (2018-01-30)
Huawei Proprietary and Confidential
Copyright © Huawei Technologies Co., Ltd.
30
