NEC Active Upgrade Express5800/320Fc, Site Planning Manual

Page: 12 / 62

Site Planning for Fault-Tolerant Systems
1-2
Express5800/320
Fc
: Site Planning Guide
Use the “Site Planning Checklist” on page 1-3 to track your site preparation progress.
For a list of other documents related to your Express5800/320Fc system, see “System
Documentation” on page 1-5 .
During the site planning and preparation processes, work closely with your facilities
group or contractor to determine space, power, and environmental requirements. Enlist
their help to provide a suitable location with sufficient alternating current (AC) power,
heating, ventilation and air conditioning (HVAC) capabilities, and network and
telephone connections.
If your system is covered by a service agreement and you need help with site planning,
contact the NEC Technical Support. If you have a contract with the NEC Technical
Support to install the system, contact them after you have prepared the installation site
and moved the system to the site. For more information about NEC Technical Support,
visit http://support.necsam.com/servers/ft.
See Appendix A, “System Specifications” for Express5800/320Fc system specifications
and the
Express5800/320 Fc : Peripherals Site Planning Guide
for specifications of
other components.
Site Planning for Fault-Tolerant Systems
Consider the following specific fault-tolerant features of Express5800/320Fc
systems for site planning:
•
Lockstep technology
means that the Express5800/320Fc systems contain redundant
hardware. The Express5800/320Fc systems contain two enclosures, each containing a
full computing environment that consists of a CPU element and an I/O element.
If a component in a CPU element malfunctions, the corresponding CPU element in
the other enclosure, which is processing the same information in
lockstep
,
continues processing without interruption. If a component in the I/O element
malfunctions, the system fails operation over to the corresponding element in the
other enclosure and continues to operate normally. The only consequences are
that the system is less fault tolerant, and any I/O throughput distributed between
the enclosures may be reduced. To restore full fault tolerance, an enclosure can
be replaced without taking the system offline.