Incorrect output

Figure 1. Overview of the incorrect output procedure

The following procedure describes each step shown in Figure 1.

1. Document the incorrect data.
   Most incorrect output problems pertain to data contained in a PIU. This can be incorrectly formatted user data, routing information that is not valid, or other types of missing or incorrect data. These types of problems are generally difficult to diagnose, as they show up only at the user locations. From the following list, choose all the symptoms that apply to your problem and add them to your problem documentation:

   • Cursor is in the wrong place or missing.
   • Format of data is incorrect.
   • Data is truncated.
   • Data is incorrect.
   • Data is missing.
   • Problem is related to an application program macroinstruction.
   • Screen is improperly formatted.
   • Session is hung.
   • Terminal is hung.
   Note: The term hung terminal means that the user is prevented by the system from entering data.
2. Determine the device type.

   If the problem is associated with a specific device type, add the device type (for example, 3277, 3278, or 3279 terminal) to your problem documentation.

3. Is it a TSO/VTAM user problem?

   If the incorrect output problem involves TSO/VTAM, go to Screen management problems.

4. Get network trace and dump output.
   Re-create the problem with the following service aids active:

   • Start the VTAM® buffer contents trace for the failing application program and terminal logical units.
   • If you have an IBM® 3710 Network Controller, start the network controller line trace. It traces information passing to and from a 3710.
   • Start the VIT with MODE=EXT. Unless APPC is involved, do not specify the APPC option. Unless you suspect that a lock is not working, do not specify the LOCK option. Do not specify the MSG option. This shows the type of requests being processed between the application program and the user or terminal, and the control information for routing, pacing, and so on, in each PIU sent in the network.
   • If the problem involves lines or devices attached to a communication controller, start the generalized PIU trace. This shows how far the PIU got within the NCP and what the PIU looked like (its control information) when it was sent to the line.
   • If you have the NetView program, you can use the session trace data to determine the requests and responses received and sent by VTAM and the other network nodes.
   • If it is available, you can use GTF CCWTRACE to trace the CCWs, I/O interruptions, and all CCW data for each Start I/O issued by the system. For more information about these traces, see the diagnostic books for your operating system.
     Note: CCW trace will not capture data for a data device for the following devices:

     • OSA-Express®
     • HiperSockets™

     I/O trace must be used for these devices. CCW trace can be used for the control devices for the above devices.
   • If it is available, you can use the VARY TCPIP,,OSAENTA command to trace the packets sent to the network or received from the network by an OSA-Express2 or later adapter. The I/O trace captures the data as it is sent between VTAM and the OSA, but the OSAENTA trace captures the packets sent to and from the PCI bus on the adapter. For more information about the OSAENTA trace, see z/OS Communications Server: IP Diagnosis Guide.
   • As soon after the problem occurs as possible, take a dump of the application program, VTAM, and TSO/TCAS. Stop all traces, and format the dump and trace output for online viewing.

   The dump is used to reference storage addresses, such as control blocks and module entry points. The trace data shows at what point the data was modified, and what PABs the data was on as it was processed by VTAM. Take the dump during the re-create, when the traces are running. A dump taken earlier might not be accurate because the terminal device might have been deactivated and reactivated. This would allocate a different set of control blocks.

   For more information on dumps and traces, see Using dumps and Using traces. Operating system service aids are documented in operating system publications.

5. Examine the trace output.

   Examine the individual trace entries to find the failure. If the problem concerns user data format, and the buffer contents trace or PIU trace does not show the incorrect data, use the output from the VIT trace with the SSCP option.

   Use GTF CCWTRACE (if available) to see whether data is correct when it is sent to the NCP or logical unit.

   Use the full buffer contents option for this trace. To use the full buffer contents option, specify AMOUNT=FULL on the buffer contents trace START option or on the MODIFY TRACE command. The VTAM internal trace records CC2, CI2, and CO2 contain the first 24 bytes of this data.

   When output data is correct:

   If the traces show that the data or the control information in the RH/TH as it leaves VTAM is correct, the problem is not in VTAM or the application program; go to step 9. If the data going to the network is not valid, continue with step 7.

   When input data is incorrect:

   If the traces show that VTAM is receiving data that is not valid from a source external to VTAM, the problem is in the network; go to step 9. If the data from the network is valid, the problem is in VTAM or an application program; continue with step 7.

6. Is it VTAM or an application program?

   The problem has been narrowed down to VTAM or the application program. Examine each trace entry to determine whether the information from the application program was incorrect. If VTAM seems to be responsible, go to Reporting the problem to IBM; otherwise, continue with step 8.

7. Is more application program help needed?

   For IBM application programs such as CICS® or IMS™, you can find additional diagnostic help in the IBM application program documentation. If you decide that the problem is with an IBM application program, contact the appropriate IBM representative for that product.

8. Is the problem with an external network device?

   The problem has been narrowed down to the VTAM network, but not to VTAM itself. Try to identify the device or program responsible. You can use service aids, such as the NCP line trace, generalized PIU trace, or transmission group trace, to trace data flow between the NCP and terminal logical units. For information about how to use these traces, see Traces provided by NCP. For OSA devices, you can use CCW or I/O trace to trace data flow between VTAM and the OSA, and you can use the OSAENTA trace to trace data flow between an OSA-Express2 or later adapter and the network. For a 3720, 3725, or 3745, use the scanner interface trace (SIT) to distinguish between NCP problems and line or terminal problems.

   If you suspect the NCP, see Table 1 to determine what document contains information on troubleshooting NCP problems.

   Using dumps explains how to use system dumps, including the NCP dump. Contact the appropriate IBM representative for the device or program identified as the cause of the problem.

9. Report the problem.

   Go to Reporting the problem to IBM.