Availability - Fujitsu SPARC Enterprise M8000 Overview Manual

Since ECC protects functional data in all routes including a computing unit, a
■
register, cache memory, and a system bus, all 1-bit errors can be automatically
corrected by hardware to ensure data integrity.
2.4.2

Availability

Availability is characterized by how easily a server fails and how quickly the user
can be recovered from the failure. The amount of time the system is usable is
represented as a percentage.
Hardware and software faults in the system cannot be completely eliminated. To
provide high availability, the system must include mechanisms that enable
continuous system operation even if a failure occurs in hardware, such as
components and devices, or in software, such as the operating system, or application
software.
SPARC Enterprise M8000/M9000 servers provide the functions listed below to
obtain high availability. Higher availability can also be obtained by combining the
server with clustering software or management software.
Supporting redundant configurations and active (hot) replacement of Power
■
Supply Units and FAN Units
Supporting redundant configuration of hard disk drive, mirroring by software
■
and active replacement
Extended range of automatic correction of temporary faults in memory, system
■
buses, and LSI internal data
Supporting an enhanced retry function and degradation function for detected
■
faults
Shortening the downtime by using automatic system reboot
■
Shortening the time taken for system startup
■
XSCF collection of fault information, and preventive maintenance using different
■
types of warnings
Supporting the Chipkill function in the memory subsystem, which enables single-
■
bit error correction to continue processing in response to continuous burst read
errors caused by failures of a memory device
Supporting the memory mirroring function enables normal data processing
■
through the other memory bus, thereby preventing system failures in response to
an error at the bus or device connected to memory bus
Memory patrol function has no influence on the workload of software operation
■
because it is implemented in hardware
Chapter 2 System Features 2-15