Table 5-4

Advanced Entry Class Troubleshooting (continued)

Step

Symptom/Condition

Action

 

 

 

8a

Machine Check Abort (MCA)

Front panel LEDs indicate that the server detected a fatal error that

 

occurs during server operation;

it cannot recover from using OS recovery routines (system health is

 

the server reboots the OS.

flashing red, internal health is steady green, external health is steady

 

 

green, and power is steady green).

 

 

1. Capture the MCA dump with the EFI command:errdumpmca .

 

 

If the system can boot the OS, you can capture binary MCA dump

 

 

files online.

 

 

2. Examine the iLO 2 MP logs for entries related to processors,

 

 

processor power modules (PPMs), shared memory, and core I/O

 

 

devices. See “Errors and Reading Error Logs” (page 154) for

 

 

information.

 

 

The problem is fixed when the MCA does not repeat, or the source

 

 

of the MCA has been determined and repaired.

 

 

 

8b

Machine Check Abort (MCA)

The front panel LEDs indicate that the server detected a fatal, front

 

occurs during server operation;

side bus error caused by MBEs reading cache or DIMM, or by any

 

server reboot of OS is prevented.

parity in the I/O path between SBA, LBA, or HBA (system health is

 

NOTE: The troubleshooting

off, internal health is flashing amber, external health is steady green,

 

and power is steady green).

 

actions for this step are identical

 

System firmware gathers and logs all error data for this MCA event.

 

to those in Step 8a, except that

1. Examine the iLO 2 MP logs for entries related to processors,

 

the server in this step must be

 

processor power modules (PPMs), shared memory, and core I/O

 

hard reset to begin the booting

 

devices. See “Errors and Reading Error Logs” (page 154) for

 

process

 

information.

 

 

 

NOTE: You must hard reset the

The problem is fixed when the MCA does not repeat.

 

server to clear the fatal condition

 

 

and boot the OS.

 

 

 

 

Troubleshooting Tools

There are a number of tools available for troubleshooting HP Integrity rx3600 servers, including LEDs, online and offline diagnostics, and system fault management tools.

LEDs

The HP Integrity rx3600 server provides front panel LED, diagnostic panel LED, and FRU health LEDs, which allow you to monitor system health.

Front Panel

The front panel of the HP Integrity rx3600 server contains the power button/System Power LED, Internal Health LED, External Health LED, System Health LED, UID Button/LED, and an INIT button. The server uses flashing states (for example, flashing amber or red) on these LEDs to indicate a warning or an error. See Figure 5-1for a detailed layout of these front panel indicators and buttons.

When you power on the server, the LEDs light up as follows:

1.Power LED turns green when the system starts to power up

2.External health indicator turns green when the BMC powers on

3.Internal health indicator turns green when the BOOT_START event is received from system firmware

NOTE: BOOT_START can be determined by LED states on the I/O backplane board.

4.The system health indicator turns green when the firmware leaves exit boot services and the server starts to boot an OS.

Troubleshooting Tools 147