Table 5-22 Cooling Subsystem Events that Light Diagnostic Panel LEDs

Diagnostic

IPMI Events

Cause

Source

Notes

LEDs

 

 

 

 

 

 

 

 

 

Fans (1-3)

Type 02h, 0Ah:07h:01h

The fan has failed or

BMC

Replace the fan.

 

COOLING_UNIT_WARNING

is degrading.

 

 

 

 

 

 

 

Fans (1-3)

Type 02h, 0Ah:07h:02h

The fan has failed

BMC

Replace the fan.

 

COOLING_UNIT_FAILURE

and no longer meets

 

 

 

 

minimum cooling

 

 

 

 

requirements.

 

 

 

 

 

 

 

Common I/O Backplane

This subsection provides information on troubleshooting issues with the eight public and two private PCI-X slots.

I/O Subsystem Behavior

The I/O subsystem transfers blocks of data and instruction words between physical shared memory and virtual memory (system disks/disk array). The system boot is the first time blocks of data and instructions words are transferred into physical shared memory from a local disk or DVD or from a remote disk on another server using multiple LAN transfers. This process is referred to as Direct Memory Access (DMA). It is initiated by I/O devices located in core I/O or on Host Bus Adapter (HBA) I/O cards and does not involve any logical CPUs.

The I/O subsystem also transfers data between the internal registers within each CPU core and the internal control registers within the Zx2/PDH/Local Bus Adapters (LBA) and HBA chips. This process is referred to as programmed I/O. It is initiated by logical CPUs executing external LOAD/STORE instructions. Both system firmware and the OS kernel use this method to initiate DMA transfers.)

I/O Messages

Use the SEL to identify low level I/O subsystem faults if LEDs on the diagnostic panel are not lit. IPMI events in the SEL and FPL provide the logical ACPI path of the suspect I/O subsystem CRU. To determine the physical PCI slot number, see Table 5-25.

For configuration errors, for example, no core I/O HBAs installed or working, see “Supported Configurations” (page 155).

Diagnostic messages are reported for high level I/O subsystem errors. All fatal I/O subsystem errors cause global MCAs. HP-UX provides its own path with the physical rope number of the suspect I/O subsystem CRU. To determine the physical PCI slot number, see Table 5-25.)

Table 5-23and Table 5-24lists the I/O card events that light and may light the diagnostic panel LEDs.

Table 5-23 I/O Card Events that Light Diagnostic Panel LEDs

Diagnostic

IPMI Events

Cause

Source

Notes

LEDs

 

 

 

 

 

 

 

 

 

I/O card

Type 02h, 03h:05h:01h

Overcurrent on PCI

BMC

Possible short circuit on

 

CURRENT_LIMIT_EXCEEDED

slot.

 

the I/O card or I/O slot.

 

 

 

 

 

I/O assembly

Type 02h, 02h:07h:03h

The voltage on the

BMC

A voltage on the I/O

 

VOLTAGE_DEGRADES_TO_NON_RECOVERABLE

CRU is inadequate.

 

backplane is out of range

 

 

 

 

(probably too low).

 

 

 

 

 

Disk

Type 02h, 02h:07h:03h

The voltage on the

BMC

A voltage on the I/O

assembly

VOLTAGE_DEGRADES_TO_NON_RECOVERABLE

CRU is inadequate.

 

backplane is out of range

 

 

 

 

(probably too low).

 

 

 

 

 

Common I/O Backplane 167