Network Topologies; Points Of Failure; Host Connection Topologies; Drawing Legends And Configuration Concerns - Infortrend EonStor S12E-R1132-4 Installation And Hardware Reference Manual

EonStor S12E-R1132-4/G1133-2 Installation and Hardware Reference Manual
All cables must be handled with care. Make sure Ethernet cables do not
run parallel with power cords. To prevent interference within a rack
system, the cable routing path must be carefully planned and they must
not be bent.
4.1.2

Network Topologies

The iSCSI host ports connect to Ethernet network devices and iSCSI
initiators that comply with the IETF iSCSI standard (RFC 3720). Network
connection of the iSCSI ports is flexible. The use of network connecting
devices, subnet, Name Servers, or iSCSI management software can vary
from case to case. In Section 4.2.2 to 4.2.5, configuration diagrams will be
provided to illustrate the possible uses of the EonStor S12E subsystem in
an iSCSI network.
Note that these configurations only recommend but do not intend to impose
limitations on the use of subnet settings or networking paths to data servers.
4.1.3

Points of Failure

The primary concern for configuring host-side topologies is to avoid points
of failure. It is therefore preferred that two HBA cards are installed on the
host side and better not the onboard Ethernet. Data flow and access
management should therefore be implemented to avoid access contention.
Infortrend's EonPath ™ software can be implemented in order to access a
RAID volume through an alternate data link in the event of cabling failure.
4.2

Host Connection Topologies

4.2.1

Drawing Legends and Configuration Concerns

In Section 4.2.2 to 4.2.5, we provide some sample connection topologies.
The following should be taken into consideration with these configurations:
․
․
․
4-2
WARNING!
Use the 10/100BaseT Ethernet port for management purposes, i.e.,
SANWatch or telnet console. This 10/100BaseT port is not used for
I/O transactions.
Configurations such as subnet or Virtual LAN can separate iSCSI
host links, decrease overheads, and eliminate the impact of link
failure.
Multiple arrays or logical partitions can be created and made
available separately through different IDs or LUN numbers on the