Chapter 12. X.25 Problem Determination

Flashing 888 Problems

Flashing 888-102

An initial value of 102 indicates an unexpected system halt during normal operation.

For unexpected system halts, the string of three-digit display values has the following format:

888 102 mmm ddd

where mmm is a value indicating the cause of the halt and ddd is a value indicating whether or not a system dump was obtained.

Refer to the hardware problem determination procedures supplied with your system. If these procedures return an SRN, record that SRN in item 4 of the Problem Summary Form and report the problem to your service organization.

If the diagnostics do not detect a problem, record SRN 101-mmm in item 4 of the Problem Summary Form and report the problem to your service organization. If a system dump was obtained, copy the dump to removable media and be prepared to make it available to your service organization.

The following list gives the possible values of mmm, the second value that follows the 888, and the cause of the system halt invoking that value:

• 200 Machine check due to memory bus error (RAS/CAS Parity).
• 201 Machine check due to memory timeout.
• 202 Machine check due to memory card failure.
• 203 Machine check due to address exception: address out of range.
• 204 Machine check due to attempted store into ROS.
• 205 Machine check due to uncorrectable ECC due to address parity.
• 206 Machine check due to uncorrectable ECC.
• 207 Machine check due to undefined error.
• 300 Data storage interrupt - processor type.
• 32x Data storage interrupt - input/out exception - IOCC. The number represented by x is the BUID.
• 38x Data storage interrupt - input/output exception - SLA. The number represented by x is the BUID.
• 400 Instruction storage interrupt.
• 500 External interrupt - Scrub - memory bus error (RAS/CAS Parity).

  External interrupt - DMA - memory bus error (RAS/CAS Parity).

  External interrupt - undefined error.

• 52x External interrupt - IOCC type - channel check.

  External interrupt - IOCC type - bus timeout.

  External interrupt - IOCC type - keyboard external.

  The number represented by x is the IOCC number.

• 700 Program interrupt.
• 800 Floating point unavailable.

The value of ddd, the third value following the 888, indicates the current dump status. The possible values and meanings of ddd are:

• 0c0 Dump completed successfully.
• 0c4 Partial dump completed.
• 0c5 Dump failed to start. An unexpected error occurred while the system was attempting to write to the dump device.
• 0c8 Dump failed. No primary dump device is configured.

Flashing 888-103

An initial value of 103 indicates a diagnostic message. Diagnostic messages are displayed in the three-digit display when the console display is not present, or is unavailable because of a display or adapter failure, or when a failure is detected that prevents the completion of IPL.

The string of three-digit display values identifies the SRN, and up to four field replacement Units (FRUs). The string of three-digit display values has the following format:

• 888 103 nnn nnn c01 1ee 2ee 3dd 4dd 5 ss 6ss 7ff 8ff
• c02 1ee 2ee 3dd 4dd 5ss 6ss 7ff 8ff
• c03 1ee 2ee 3dd 4dd 5ss 6ss 7ff 8ff
• c04 1ee 2ee 3dd 4dd 5ss 6ss 7ff 8ff

The two values nnn nnn represent the SRN. The values c01, c02, c03 and c04 indicate the first, second, third and fourth FRUs, respectively. For each FRU, the value sequence 1ee 2ee 3dd 4dd 5ss 6ss 7ff 8ff is the location code. Refer to your diagnostic information for interpretation of these location codes.

Record the SRN in item 4 of the Problem Summary Form and the location codes in item 6 of the Problem Summary Form. Then, report the problem to your service organization.

Forcing a System Dump

If the system did not produce a dump automatically because of a hang condition, obtain a dump while the problem exists.

Note: The use of the Reset button is the preferred method of obtaining dumps, because in a hang condition it is the dump trigger most likely to succeed. If the system has dumped and halted automatically, the Reset button will scroll the LEDs rather than trigger another dump.

Attention: Obtaining a dump overwrites a previous dump or other data stored on the dump device.

If the console or a tty is accepting commands, you can start a dump using the sysdumpstart command. The sysdumpstart command allows you to start a dump to the primary or the secondary dump device. If the system is accepting commands, use the following procedure:

1. To determine which devices have been assigned as the primary and secondary dump devices, enter:

   sysdumpdev -l

2. To start a dump to the primary dump device, log into the system as the root user and enter:

   sysdumpstart -p

If the system is not accepting commands, then try one of the following:

• If there is a keyboard attached to the system unit, you can start a dump using the dump key sequences (Ctrl-alt-numpad1 and Ctrl-alt-numpad2). The dump key sequences allow you to start a dump to the primary or the secondary dump device.
• The Reset button can be used to start a dump to the primary dump device. To start a dump, turn the Key Mode switch to the Service position and press the Reset button.

X.25 Problem Diagnosis

Before investigating any problem, ensure that X.25 communications are set up correctly. The following commands may help diagnose the problem:

• The x25mon command.
• The lsx25 command.
• X.25 clear and reset codes .

If it is required to trace the internal working of the X.25 stack, the following trace points are available. These traces, along with a line trace from x25mon, a microcode trace using sx25debug and a system configuration table form lsx25, would help diagnose the code's behavior.

The traces produced from these trace points are not in a form that is directly useful to a system user or system administrator. They are designed to allow diagnosis of code flow.

The frame layer and physical layer code, running on the adapter, are also equipped with error point tracing. If necessary, it is possible to capture the error logs from the microcode using sx25debug.

See the following for further discussion of X.25 problems and solutions:

• Diagnosing Problems with Connecting to the X.25 Network
• Diagnosing Problems with Making an Outgoing X.25 Call
• Diagnosing Problems with Receiving an Incoming X.25 Call
• Diagnosing X.25 Packet Problems
• Diagnosing X.25 Command Problems.

Starting Traces

The following procedure describes how to take a trace:

1. Start the traces (x25mon and system traces) using the trace command:

   trace -a -j 25C, 33B, 329, 33C

2. Recreate the problem.
3. Stop the trace using the following commands:

   trcstop 
    
   kill x25mon

Possible Causes

1. Verify that the cables are attached correctly and are secure.
2. Verify that the modem-to-modem eliminator/switch is working properly.
3. Verify that the board is seated properly.
4. Verify that the device driver and ports are configured correctly.

Diagnosing Problems with Connecting to the X.25 Network

Problem 1

When adding a port or driver, the add fails with the message A device is already configured at the specified location .

Suggestions

• This indicates that another device is using the resources that are needed. For example, an X.25 port is already configured on the port number requested. Another example is a device driver is already available on the adapter selected. The lsx25 command shows how the X.25 system is configured.

Problem 2

Attempts to connect an X.25 port fail when making the port via mksx25 or through SMIT. The physical and the frame layers remain down. The x25mon command shows that there is no line activity.

Suggestions

• The cable is absent or loose.
• There is no carrier (DCD or RLSD) from the network terminating unit (NTU).
• There is a cabling problem such as an extra null.
• The X.25 adapter is not seated correctly.

Problem 3

Attempts to connect an X.25 port fail. The physical layer is established for several seconds, as shown by the lights on the NTU, but then goes down again. The x25mon command indicates that there is no line activity.

Suggestions

• The X.25 adapter is expecting a clock signal and not receiving one. Adjust the NTU to provide clocking.
• The cable is loose, causing the clock pin not to be connected.

Problem 4

Attempts to connect an X.25 port fail. The physical layer is connected but the frame layer fails to come up. x25mon monitoring the frame layer shows a sequence such as a string of SABM s.

Suggestion

The packet layer type of line attribute is DCE rather than DTE for this X.25 adapter. Ensure that the DTE/DCE switching configured in SMIT is suitable for the device being attached to.

The frame layer might need to be set for automatic detection. Use SMIT to change it.

Diagnosing Problems with Making an Outgoing X.25 Call

Problem 5

On starting xtalk, the initial screen contains the message You cannot make outgoing calls . This implies that there are no COMIO emulation ports, x25s .

Suggestion

Use lsx25 to see if there any COMIO emulation ports available. If not, add COMIO emulation to any of the X.25 ports, sx25a , that require them.

Problem 6

Incoming calls are arriving, but outgoing calls cannot be made, on a switched virtual circuit (SVC).

Suggestions

• The configuration of the SVC logical channel numbers is incorrect. Check your network subscription and use SMIT to check the network attributes.
• Use SMIT to check that outgoing calls and local charges are allowed.

Problem 7

A call is being cleared with a cause code from 1 through 127.

Suggestion

The diagnostic code gives details of why the call is being cleared. Common causes are that the network user address (NUA) is not known, the X.25 line is not connected, or unsupported optional facilities have been requested.

Note: Cause codes 80 through FF are set by SNA. When SNA is being used, the diagnostics are interpreted differently.

Diagnosing Problems with Receiving an Incoming X.25 Call

Problem 8

The x25mon command indicates that an incoming call has arrived at the adapter, but it is not being routed to the application that is running.

Suggestions

• If the call is being cleared with cause 0 and diagnostic 0, it may be that the application is not listening to a name in the routing list that matches the incoming call. There may be another name in the routing list that is a better match for the call and has reject (R ) specified as the action.
• The incoming call may have requested an optional facility that is not allowed by the current configuration.
• The incoming call may have arrived on an invalid logical channel. Check your network subscription and the network configuration attributes in your SMIT configuration.
• Use SMIT to check that incoming calls are allowed.
• Check the cause and diagnostic codes with the standard list of codes and any network-specific codes.
• A different application may be listening to criteria that are a better match those which your application is listening. Check for another program using X.25.

Diagnosing X.25 Packet Problems

Problem 9

Whenever a packet is sent on a permanent virtual circuit (PVC), the network sends a clear-indication packet.

Suggestion

The PVC logical channel number ranges are set incorrectly. Check your network subscription and your SMIT configuration.

Problem 10

When sending a data packet with the D-bit set on an switched virtual circuit (SVC), a reset-request packet is sent instead.

Suggestion

The intention to use D-bits must be made clear when the call is originally established, either in the call-request packet or in the call-accepted packet.

Problem 11

When sending a data packet with the D-bit set on a permanent virtual circuit (PVC), a reset packet is sent instead.

Suggestion

Use SMIT to configure the PVC to use D-bits.

Problem 12

When sending an interrupt packet with more than one byte of user data, a reset packet is sent instead.

Suggestion

The 1980 version of X.25 supports exactly one byte of user data in the interrupt packet. The 1984 version supports up to 32 bytes. Check your subscription and the value of the CCITT support attribute in SMIT.

Problem 13

Suggestion

The CCITT timer, T1, may have expired. Use SMIT to increase the value of the T1 attribute, or use smaller packets. Before changing T1 or packet size values, verify these values with your network provider.

Diagnosing X.25 Command Problems

Problem 14

Attempts to start up the x25mon command on the X.25 port fails.

Suggestion

Only root may start the X.25 monitoring.

Problem 15

The xmanage, xcomms and xmonitor commands are not in AIX Version 4.

Suggestion

These commands are not shipped with the V4 LPP. The x25mon command replaces xmonitor, and the x25mng sample helps provide link status on ports that have COMIO emulation.

Problem 16

None of the X.25 commands get past the title screen.

Suggestion

Use the echo $TERM command to find out your terminal-type setting and make sure that it matches your actual terminal type.

Problem 17

The number of virtual circuits cannot be configured past an upper limit.

Suggestion

The maximum number of VCs supported depends on the license of the product that was purchased. Use smit chg_sx25vcs to see the licensed number of VCs.

Diagnosing PAD Problems

Unable to get a CALL to complete after running xspad -l sx25a0. For example:

CLEAR DTE 0 136 - 
Call cleared, by remote device, data may be lost
 

Verify that the side you are calling has a PAD or X.29 daemon running. Verify that the switch/modem-to-modem/eliminator is configured and set up to acknowledge your call. Also verify that the NUA is configured properly.

When logging in through the PAD, the display isn't working as expected. For example, no backspace, echo, problems with row and column widths, etc.

Check the sttty settings and verify that they are set like a ASCII terminal. Verify that your echo, row, and columns are set correctly.

Unable to get xspad -l sx25a0 to work. For example:

Invalid port 'sx25a0

Verify that port sx25a0 is configured.