64Chapter 8 Storage enclosure management
l
Host ports used for replication must use the same protocol as the connected controllers: either FC or iSCSI.
l
You cannot directly attach both systems. System controllers require connection through switches to the
same fabric or network, whether the secondary system is remotely located or co-located with the primary
system.
l
The two associated standard volumes form a replication set, where you can only access the primary or
source volume.
l
Connection is through the primary or source volume, but if the primary system goes offline, you can access
the replicated date from the connected secondary system.
l
You must provide for a sufficient number of ports to perform replication. The primary system needs to
balance the load across those ports as input and output demands rise and fall. If some of the replicated
volumes connect to Controller 0A, and others to Controller 0B, you need at least one port on each controller
dedicated exclusively to replication. If you have a heavy replication traffic load, you need more.
l
For remote replication instances, you must confirm that all ports designated and assigned for replication
can properly communicate with the secondary system using the CLI command: query peer-connection.
l
The storage system does not provide for specific assignment of ports for replication. However, you can
accomplish a similar function by using either iSCSI virtual LANs (VLANs) or FC zones or by using
physically separate infrastructure.
l
For system security, do not expose network ports on controller enclosures to external network connections.
Cabling for replication requires planning and thought. The samples included in this section are not exhaustive,
but should serve as an optimal, conceptual guideline for your custom system configuration.
Co-located replication configurations
Volume replication can occur on the same physical network, or on different physical networks. For situations
where you need to use a single physical network, you can logically separate networks by configuring fibre
channel (FC) zones or iSCSI virtual LANs (VLANs) to isolate input and output (I/O) traffic from replication
traffic. In such a case, while such configurations physically inhabit a single network, they logically function as
multiple servers and networks.
Co-located replication with multiple switches and servers
Ideally, use three separate switches to avoid a single point of failure and allow for physical isolation of I/O
traffic from replication traffic. Two switches are dedicated to I/O traffic and act as the bridge to connect
controller enclosures to multiple host servers. The third switch is the replication switch and acts as the bridge
to connect controller enclosures to each other.
Sample cabling for the first controller enclosure:
l
Two SFP I/O cables connect Controller 0A and two more connect Controller 0B to the left I/O switch.
l
Two SFP replication cables connect Controller 0A and two more connect Controller 0B to the center
replication switch.
Sample cabling for the second controller enclosure:
l
Two SFP I/O cables connect Controller 0A and two more connect Controller 0B to the right I/O switch.
l
Two SFP replication cables connect Controller 0A and two more connect Controller 0B to the center
replication switch.
To Next Page
Loading...
Need help?
General
