Infortrend
8-2
8.2
Major Concerns
Most of the configuration options in this chapter are directly related
to controller redundancy. Joining two controllers into a dual-active
pair can eliminate most possible points of failure. Configuring a
controller pair requires careful planning and proper setup and the
requirements can be summarized as follows:
!
Redundant Cache Coherency Channels (RCC):
1.
RCC
FC channels can be manually assigned as
the dedicated communications loops, two
are recommended for path redundancy and
sufficient bandwidth.
2. Drive + RCC
Communications traffic distributed over
drive loops
!
Connection between Controllers:
Cabling between controllers, hardware link through a
common backplane, Fibre hub or switch (for SAN
applications and for those models that do not have by-
pass chips)
!
Channel Mode Assignment
According to the topological plan, your I/O channels
can be designated as:
•
Host
•
RCC paths
•
Drive
•
Drive + RCC
!
Host Channel Connection Type:
This depends on the way your RAID system is
connected to the host computer(s). The host
connection type can be:
•
FC-AL
•
Fabric (point-to-point)
!
Controller Unique ID:
This ID will be used to generate Fibre ports’ node
names, and is necessary for addressing the controller
during the controller failover/failback operation.
!
Dual-Loop:
1.
Drive-side
dual loop provides data path
redundancy. Firmware is capable of executing a
load-sharing algorithm to optimize dual-loop
