Performing Remote Host System Diagnostics

C H A P T E R 15

Topics
Description	Links
Diagnose x86 system hardware issues	Diagnosing x86 Systems Hardware Issues
Diagnose SPARC system hardware issues	Diagnosing SPARC Systems Hardware Issues
Collect data for use by Oracle Services personnel to diagnose system problems	Collecting SP Data to Diagnose System Problems

Related Topics
For ILOM	Chapter or Section	Guide
Concepts	Diagnostics for x86 or SPARC Systems Collect SP Data to Diagnose System Problems	Oracle Integrated Lights Out Manager (ILOM) 3.0 Concepts Guide (820-6410)
Web interface	Diagnostics Collect SP Data to Diagnose System Problems	Oracle Integrated Lights Out Manager (ILOM) 3.0 Web Interface Procedures Guide (820-6411)
The ILOM 3.0 Documentation Collection is available at: `http://docs.sun.com/app/docs/prod/int.lights.mgr30#hic`.

Diagnosing x86 Systems Hardware Issues

Topics
Description	Links	Platform Feature Support
Review the prerequisites	Before You Begin	x86 system server SP
Ensure that the requirements for configuring and running diagnostic tests are met	Configure and Run Pc-Check Diagnostics


Configure and run Pc-Check diagnostic tests	Configure and Run Pc-Check Diagnostics
Generate a NMI to a host	Generate a Non-Maskable Interrupt
Run other x86 system hardware diagnostic tests and tools	Sun x64 Servers Diagnostics Guide (820-6750)

Before You Begin

To diagnose x86 systems hardware issues, you need the Reset and Host Control (r) role enabled.

Configure and Run Pc-Check Diagnostics

1. Log in to the ILOM SP CLI.

2. Type the following commands to enable the diagnostic tests:

-> cd /HOST/diag/
/HOST/diag
 
-> show /HOST/diag
   Targets:
 
   Properties:
	 	state = disabled
 
   Commands:
       cd
       set
       show
 
-> set state=extended 	This will enable Pc-Check to run a 20-40 minute test suite
OR 
-> set state=enabled 		 	 This will enable Pc-Check to run a 4-5 minute test suite
OR
-> set state=manual 		 	 This will enable you to select specific Pc-Check tests to run
 
-> show
   Targets:
 
   Properties:
       state = enabled
 
   Commands:
       cd
       set
       show

3. Reset the power on the host to run the PC diagnostic tests.

Generate a Non-Maskable Interrupt

Caution - Depending on the host OS configuration, generating a non-maskable interrupt (NMI) might cause the OS to crash, stop responding, or wait for external debugger input.

1. Log in to the ILOM SP CLI.

2. Type the following commands:

-> cd /HOST 
/HOST
 
-> show
/HOST
Targets:
		diag
 
Properties:
		generate_host_nmi = (Cannot show property)
 
Commands:
		cd
		set
		show
 
-> set generate_host_nmi=true
set ‘generate_host_nmi’ to ‘true’

Diagnosing SPARC Systems Hardware Issues

Topics
Description	Links	Platform Feature Support
Review the prerequisites	Before You Begin	SPARC system server SP
Configure the system to run diagnostic tests	Configure Diagnostics Mode


Specify which diagnostic triggers to activate	Specify the Diagnostics Trigger
Specify the level of diagnostics that you want to execute	Specify Level of Diagnostics
Specify the verbosity output of the executed diagnostic tests	Specify Verbosity of Diagnostics Output

Before You Begin

Prior to performing the procedures in this section, the following requirement must be met:

To configure and run diagnostic tests on a SPARC system, you need the Reset and Host Control (r) role enabled.

Configure Diagnostics Mode

Use the /HOST/diag host mode property to control whether diagnostics are enabled and to specify which diagnostic mode is enabled.

Follow these steps to configure the diagnostic mode:

1. Log in to the ILOM SP CLI.

2. At the command prompt, type the following command:

-> set /HOST/diag mode=value

Where value is one of the following:

off - Do not run any diagnostics.

normal - Run diagnostics (the default value).

3. Reset the power on the host to run the diagnostic tests.

Specify the Diagnostics Trigger

You can select one or more triggers that will cause a power-on self-test (POST) to be run on the host.

Follow these steps to set the trigger levels:

1. Log in to the ILOM SP CLI.

2. At the command prompt, type the following command

-> set /HOST/diag trigger=value

Where value can be one of the following:

none - Diagnostics will not be triggered to run.

user-reset - Diagnostics will be run upon a user-invoked reset.

power-on-reset - Diagnostics will be run when power is applied.

error-reset - Diagnostics will be run upon any error-invoked reset.

all-resets - Diagnostics will be run for any of the above reset types.

Specify Level of Diagnostics

There are separate ILOM CLI properties that enable you to specify the level of diagnostic testing to be executed, depending on how the diagnostics were triggered to run. This gives granular control of how much diagnostic testing is performed in different host reset situations.

Use the /HOST/diag level property to specify the level of diagnostic testing to be executed when diagnostics are enabled.

Follow these steps to specify the level of diagnostics to be executed:

1. Log in to the ILOM SP CLI.

2. Perform the one of the following commands, depending on how the host is reset:

To specify the diagnostic level when the host is powered on, type the following command:

> set /HOST/diag power_on_level=value

To specify the diagnostic level when the host is reset by the user, type the following command:

-> set /HOST/diag user_reset_level=value

To specify the diagnostic level when the host is reset due to a system error, type the following command:

-> set /HOST/diag error_reset_level=value

Where value is one of the following:

min - Run the minimum level of diagnostics to verify the system.

max - Run the maximum set of diagnostics to fully verify system health (the default value).

Note - For backward compatibility with ILOM 2.x, the former property /HOST/diag level is still supported as a shortcut for specifying the same diagnostic level for all trigger types. Any value set to /HOST/diag level will be applied to all three trigger-specific properties: power_on_level, user_reset_level, and error_reset_level.

3. Reset the power on the host to run the diagnostic tests.

Specify Verbosity of Diagnostics Output

There are specific ILOM CLI properties that enable you to specify the output verbosity of executed diagnostics, depending on how the diagnostics were triggered to run. This gives granular control of how much diagnostics output is given in different host reset situations.

Follow these steps to specify the verbosity of the diagnostics output:

1. Log in to the ILOM SP CLI.

2. Perform one of the following commands, depending on how the host was reset:

To specify the output verbosity for diagnostics executed when the host is powered on, type the following command:

> set /HOST/diag power_on_verbosity=value

To specify the output verbosity for diagnostics executed when the host is reset by the user, type the following command:

-> set /HOST/diag user_reset_verbosity=value

To specify the output verbosity for diagnostics executed when the host is reset due to a system error, type the following command:

-> set /HOST/diag error_reset_verbosity=value

Where value is one of the following:

none - Diagnostics do not print any output on the system console when running, unless a fault is detected.

min - Diagnostics print a limited amount of output on the system console.

normal - Diagnostics print a moderate amount of output on the system console (the default value).

max - Diagnostics print full output on the system console, including the name and results of each test being run.

debug - Diagnostics print extensive debugging output on the system console, including devices being tested and debug output of each test.

Note - For backward compatibility with ILOM 2.x, the former property /HOST/diag verbosity is still supported as a shortcut for specifying the same output verbosity for all trigger types. Any value set to /HOST/diag verbosity will be applied to all three trigger-specific verbosity properties: power_on_verbosity, user_reset_verbosity, and error_reset_verbosity.

3. Reset the power on the host to run the diagnostic tests.

Collecting SP Data to Diagnose System Problems

Topics
Description	Links	Platform Feature Support
Review the prerequisites	Before You Begin	Oracle Service personnel feature only
Collect SP data	Collect SP Data to Diagnose System Problems

Before You Begin

To collect SP data using the Service Snapshot utility, you need the Admin (a) role enabled.

Caution - The purpose of the ILOM Service Snapshot utility is to collect data for use by Oracle Services personnel to diagnose problems. Customers should not run this utility unless requested to do so by Oracle Services.

Collect SP Data to Diagnose System Problems

Follow these steps to run the Service Snapshot utility:

1. Log in to the ILOM SP CLI.

2. Type the following commands:

->set /SP/diag/snapshot dataset=data->set /SP/diag/snapshot dump_uri=URI

Where data and URI are one of the following:

Variable	Option	Description
data	`normal`	Specifies that ILOM, operating system, and hardware information is to be collected.
	FRUID	Available as of ILOM 3.0.3, requests ILOM to collect information about FRUs currently configured on your server in addition to the data collected by the `normal` option.
	`full`	Specifies that all data is to be collected ("full" collection). Note - Using this option might reset the running host.
	`normal-logonly` `fruid-logonly` `full-logonly`	Specifies that only log files are to be collected.
URI	Any valid target directory location	Specifies the URI of the target directory. The URI format is as follows: `protocol://username:password@host/directory` Where protocol can be one of these transfer methods: SFTP or FTP. For example, to store the snapshot information in the directory named `data` on the host, define the URI as follows: `ftp://joe:`mypasswd@host_ip_address`/data` The directory `data` is relative to the user’s login, so the directory would probably be `/home/joe/data`.