TABLE 1-1Software Services (Continued)

ServiceDescription

Fault management

Hot-replacement operations

External I/O Expansion Unit management

No initial configuration is needed.

•Domain fault management includes CPU, memory, and I/O (PCI/PCIe) nonfatal errors. All nonfatal errors are reported to the Solaris OS, which will attempt to take faulty CPUs offline or to retire faulty memory pages. Fatal errors are generally handled by the Service Processor.

•Service Processor fault management includes fatal CPU, memory, and I/O errors (the Service Processor will exclude the faulty components upon reboot), as well as environmental monitoring (power supplies, fan speeds, temperatures, currents) and the External I/O Expansion Unit.

Refer to the Solaris OS documentation collection.

No initial configuration is needed.

PCI cards can be removed and inserted while your server continues to operate. The Solaris OS cfgadm command is used to unconfigure and disconnect a PCI card.

Refer to the Service Manual; Solaris OS documentation collection.

No initial configuration is needed.

The External I/O Expansion Unit is a rack mountable PCI card chassis.

Refer to the External I/O Expansion Unit Installation and Service Manual.

Preparing for System Configuration

This section lists the information needed for initial system configuration and the initial configuration tasks.

Information Needed

Before you configure the software, have the following available:

■Access to the Service Processor with the appropriate privileges for your tasks. More information about access is contained in Chapter 2.

■An unused range of IP addresses for the internal DSCP network between the Service Processor and the domains.

■Network configuration information for the Service Processor, including IP addresses, netmask, DNS server, default route, NFS server.

4 SPARC Enterprise Mx000 Servers Administration Guide • November 2007