Troubleshooting the Initial Hardware Configuration

Your ASM-CS went through extensive testing and burn-in before leaving the factory; however, if your system appears to have problems starting up, use the information in this chapter to help identify the problem.

This chapter includes the following:

Troubleshooting
- Problem solving
- Troubleshooting systems
Running bootstrap diagnostic tests

Troubleshooting Overview

Typically, any problems encountered at start up will be related to the ASM-CS or to the network. Use the information in this chapter to help isolate problems and to rule out the ASM-CS system as the source of the problem.

Note Whether or not you can locate the problem, contact your service representative for information on how to proceed in resolving the problem you are having.

Before you call your service representative, have the following information ready:

Chassis type and serial number
Maintenance agreement or warranty information
Type of software and version number
Date you received the new chassis
Brief description of the problem you are having
Brief explanation of what steps you have taken to isolate the problem

Problem Solving

The key to problem solving in a system such as the ASM-CS is to try to isolate the problem to a specific subsystem. By comparing what the system is doing to what it should be doing, the task of isolating a problem is greatly simplified.

Consider the following subsystems of the ASM-CS system:

Power system--This subsystem includes the power supply, the multibus backplane and wiring, and the AC switch (circuit breaker).
Cooling system--The blower assembly should come on when power is applied and will stay on even if the power supply fails. The blower assembly is connected directly to the AC source.
System cards--This subsystem is the most difficult to troubleshoot. The light emitting diodes (LEDs) on the cards and appliques can be used to help identify a failure.

Note For complete information on interface card and applique LED indicators, refer to Appendix B "Reading LED Indicators."

System cables--This subsystem includes all of the internal and external cables that interconnect the cards and their appliques, and connect the ASM-CS to the network.

The following troubleshooting sections discuss these subsystems in more detail.

Troubleshooting the Power and Cooling Systems

Check the following items to help isolate the problem:

With the AC switch on, does the blower operate?
- If yes, the AC input checks out.
- If no, suspect the AC input, AC source, ASM-CS circuit breaker, the blower, or the power cable.
With the AC switch on and system LEDs lit, does the blower operate?
- If no, suspect the blower.
Does the system shut down after a being on short time?
- Check the environmental requirements in Chapter 2, "Preparing the Installation," and make certain that the chassis intake and exhaust vents are clear.
- Suspect a power supply failure.
System partially boots and LEDs do not light.
- Suspect a +12 and/or +5 volt power supply failure.

Troubleshooting the System Cards and Cables

Check the following items to help isolate the problem:

Card is not recognized by the system.
- Check the card connection to the backplane.
- Check the LEDs on the card and applique.
Card is recognized, but interface port(s) will not initialize.
- Check the internal interface cables for connection.
- Check the external cables for connection.
System will not boot properly or constantly/intermittently reboots.
- Check the processor card connection to the backplane and the console cable.
- Check the LED indicators (see Appendix B).
- Suspect the processor card or software.
System boots, but console screen is frozen.
- Check the external console connection.
- Verify the console baud rate in the terminal's documentation.
System powers on and boots with a particular card disconnected from the backplane.
- Suspect the card.
System powers on and boots with a particular applique or cable disconnected
- Suspect the applique or cable.

System Bootstrap Diagnostics

This section describes how to test for problems with system memory using the bootstrap program. This program can help you isolate or rule out hardware problems encountered when installing your chassis. A summary of the bootstrap diagnostic tests and command options is provided.

Diagnostic Tests

The diagnostic tests for the ASM-CS help initialize the processor hardware and boot the main operating system software.

Caution Before proceeding, read through this procedure and the "Processor Card Configuration Register" section in Chapter 5. If you have any doubts about your ability to perform any part, contact your service representative for information on how to proceed.

If you remove the jumpers from the boot file number field (bits 3, 2, 1, and 0) of the configuration register, you can start the ASM-CS in stand-alone bootstrap mode. The bootstrap mode prompt is > (angle bracket).

Once in bootstrap mode, enter ? to display a list of commands and options available to you. Following is an example of a display from a system with a CSC/4 processor card. Displays from systems with a CSC/3 processor card may vary slightly.


> ?
$ state      Toggle cache state (? for help)
B [filename] Bootload filename and start it
C [address]  Continue [optional address]
D /S M L V   Deposit value V of size S into location L with modifier M
E /S M L     Examine location L with size S with modifier M
G [address]  Start up execution
H            Offer help with commands
I            Initialize
K            Stack trace
L filename   Bootload filename, but do not start it
O            Show configuration register option settings
P            Set break point
S            Single step next instruction
T function   Test device (? for help)
Deposit and Examine sizes may be B (byte), L (long) or S (short).
Modifiers may be R (register) or S (byte swap).

Following are some of the most useful bootstrap commands:

The H command prints a summary of the bootstrap commands.
The I command causes the bootstrap program to reinitialize the hardware and clear the contents of memory.
The K command runs a stack trace on system hardware.

The B command with no argument boots the default software from ROM, assuming there are no jumpers in the boot field portion of the configuration register (bit 0). You may include an argument, filename, to specify a file to be booted over the network using TFTP. You can also include a second argument, host, which is the Internet address or name of a particular server host.
The T command runs various diagnostic tests.

: Enter T ? to display a list of the diagnostic tests:


M   Memory test
P   Probe IO/memory space

Use the memory test command to test memory. By default, the test examines on-board memory. Use the Probe I/O memory space test to look at the mapped I/O space in main memory and test for interface card responses. The test starts at address 2000000. You can specify probe increments in order to look for specific regions. The test uses the probe increments to search for registers that can be read from and written to (see the samples following).

Running the Diagnostic Tests

Follow these steps to run the bootstrap diagnostics:

Step 1: Turn off the ASM-CS and attach appropriate ESD protection.

Step 2: Remove the front access cover following steps 1-4 in the section "Chassis Access Procedure" in Chapter 5.

Note Before proceeding, read the section " Processor Card Configuration Register" in Chapter 5 to familiarize yourself with the configuration register on the processor card.

Step 3: Remove the jumpers from the boot file number field (bits 3, 2, 1, and 0) of the processor configuration register. (Note jumper positions before removal.)

Note If you start the server with the Break disable (bit 8) jumper removed from the configuration register, you can press the Break key on the console terminal to force the server into bootstrap mode. Enter c to continue normal execution of the system software.

Step 4: Restart the system. Wait for the system to print the two-line banner message and prompt you with an angle bracket (>).

Memory/Bus Diagnostic

To test memory, enter the following and then press the Return key.


>t m

Press the Return key after each prompt displayed to use the default addresses and select the default tests. Following is a sample of the Memory/Bus Diagnostic on the CSC/4 processor card:


Memory/Bus diagnostic
Starting Address [1000]?
Ending Address [400000]?
Hex argument for variable tests [FFFF]?
Select Tests [all]?
Number of passes to run [2]?
Message Level (0=silence, 1=summary, 2=normal)[2]?
Testing addresses between 0x1000 and 0x400000
Begin pass 0, test 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15   End pass
Begin pass 1, test 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15   End pass
No errors during 2 passes

Note All values are in hexadecimal; do not use the prefix 0x when entering a number.

Running a single pass of the diagnostic takes only a few minutes. If the program encounters memory problems, it displays appropriate error messages on the console terminal.

I/O Memory Space Probe

To test I/O space, enter the following and press the Return key.


>t p

The probe begins at address 2000000. You can specify probe increments. The following example shows a test done in probe increments of 10:


IO/memory space probe
Starting address [2000000]?
Ending address [210FFFF]?
Probe increment (in shorts) [1]? 10
Probing from 0x2000000 to 0x210FFFF with interval 0x10
Region 0x210C000 to 0x210C060 exists (0K)
Done
>

Table 1-1 lists the times taken to run the memory tests.

Memory Test and Memory Probe--Test Times

		Memory Probe Test Times
ProcessorCard	Memory Test Times	Probe Increment = 1	Probe Increment¹ = 10
CSC/4	~11 minutes²	~20 seconds	~2 seconds
CSC/3	~8 minutes	~30 seconds	~1 seconds

¹ In the probe test, increasing the value of the probe increment to 10 will decrease the time taken for the test, but will test in larger blocks of memory.
² All time values are aproximate (~).

This example shows a test done in probe increments of 1:


IO/memory space probe
Starting address [2000000]?
Ending address [210FFFF]?
Probe increment (in shorts) [1]?
Probing from 0x2000000 to 0x210FFFF with interval 0x1
Region 0x210C000 to 0x210C07E exists (0K)
Done
>

Table of Contents