Infortrend
10-4
The basic configuration rules are:
1.
All channels should be connected to both controllers as diagrammed
above.
2.
To reduce the chance of downtime, more than one hub or switch can
be used to connect to host computer for path redundancy.
3.
For the Fibre-to-Fibre controllers or RAID systems, there are two
options with configuring the communications loops between
controllers:
1). Dedicated Communications Loops
– “RCC”
The first option is choosing one or two Fibre loops as the
dedicated communications paths. Two for communications is
recommended for the path redundancy it provides.
Using two channels for the communications offers a greater
throughput and hence a better performance.
2). Communications over Drive Loops
– “Drive + RCC”
Configure all drive loops into the “Drive + RCC” mode to let
them share the communications traffic. The controllers can
automatically distribute the communications traffic across all
drive loops.
Workflow is balanced among loops. Using the drive/RCC mode
allows more channels to be used for drive connection. With a 6-
channel controller, for instance, there can be as many as two
channels for host and four channels for drive (drive + RCC). All
channels are used for IO traffic while the system is still benefited
from controller communications.
10.1.3 Grouping Hard Drives and LUN Mapping
Listed below are the array settings that need to be considered when
configuring a dual-controller system:
1.
H
ow many logical drives, logical volumes, or logical partitions, and
in what sizes?
2.
S
ystem drive mapping (primary/secondary ID): how many storage
volumes will appear to which host port? and managed by which
controller?
