If a failure occurs, the faulty component is automatically isolated from the system, and the system is rebooted. If 1-bit errors occur frequently in the cache memory configuring a CPU, the faulty memory can be dynamically isolated without rebooting the Solaris Operating System. This type of graceful degradation function enables the operation of the other resources to continue without interruption, and also provides high fault-tolerance in case of failure.

■eXtended System Control Facility (XSCF)

SPARC Enterprise M8000/M9000 servers use service processor called the eXtended System Control Facility (XSCF), which monitors abnormalities in hardware conditions, the operating status of domains, and system conditions such as the temperature, power supply, and fan operation.

You can configure the system to selectively degrade a faulty component for operation if an error is detected.

Scheduling is supported to enable automatic power-on and power-off of the SPARC Enterprise system according to the specified operation schedule.

The console of each domain can be controlled from the XSCF via a network.

A browser-based user interface (BUI) and the command line interface (CLI) facilitate operations for making configuration changes and status monitoring in the system.

Note – A console display terminal is required for console control. Prepare it before installation. The devices that can be used as the terminal are listed below.

■PC

■Workstation

■ASCII terminal

■Terminal server (or a touch panel connected to a terminal server)

Note – For the console connection method, see the SPARC Enterprise M8000/M9000 Servers Installation Guide.

■Partitioning function

One high-end server can be divided into multiple areas, or domains, for more effective scalability. Each domain manages resources in linkage with the XSCF. A domain may consist of optimized resources depending on its intended use, enabling more efficient system configurations.

Dynamic Reconfiguration (DR) enables adding, deleting, and relocating resources of domains without stopping processing in the domain. This enables dynamic reconfiguration of resources without stopping a job, even when the job load increases suddenly or when a faulty component is replaced.

For details on domain functions, see the SPARC Enterprise

1-6SPARC Enterprise M8000/M9000 Servers Overview Guide • April 2007