Troubleshooting Tools

• the APPC subsystem internal trace

• APPC subsystem logging

• APPC subsystem messages and return codes

• the APPCCONTROL STATUS command

• the SNACONTROL STATUS command

APPC Internal Tracing

Internal tracing can be turned on at subsystem startup or during run time. It traces the internal events of the APPC subsystem. Once the trace has been turned on, it remains on until the APPC subsystem shuts down or until you issue the APPCCONTROL TRACEOFF command.

At subsystem startup, internal tracing is turned off by default, but you can turn tracing on when you issue the APPCCONTROL START command by specifying TRACEON=YES. If you activate the APPC subsystem programmatically, you can turn tracing on by setting the TraceOn parameter of the APPCStart intrinsic to 1.

During run time, you can turn internal tracing on by issuing the APPCCONTROL TRACEON command.

	NOTE: Hewlett-Packard recommends that you turn on internal tracing when you install an LU 6.2 product on your system for the first time.

It is not necessary for you to read and interpret the internal trace, but you may be asked to collect it for analysis by your HP representative.

When you turn on APPC subsystem internal tracing, you can specify a name for the trace file, or you can allow the trace file name to default. The default trace file name is NMTCxxxx.PUB.SYS, where xxxx is a number from 0000 through 9999. This number is incremented whenever a new trace file is created.

A new trace file is created every time APPC subsystem internal tracing is turned on, at subsystem startup or during run time. To start a new trace file during run time, turn tracing off with the APPCCONTROL TRACEOFF command, and then turn it back on again with the APPCCONTROL TRACEON command.

	NOTE: When storing trace files to tape for analysis, be sure to include all the files on disk at the time the problem occurred. An incomplete set of trace files can hinder problem diagnosis.

When you turn on APPC subsystem internal tracing, you can specify a size for the trace file, in 128-word records. The trace file size can range from 0 through 32767 records. When you are trying to diagnose a problem, you should specify a trace file size of 32000 records or larger. Once a trace file has reached its specified size, trace records are overwritten starting from the beginning of the file.

APPC Logging

Node Management Services (NMS) provides the logging services for the APPC subsystem. APPC subsystem logging records the major events of the APPC subsystem.

Logging is turned on through NMMGR configuration. The APPC subsystem number is SUB0016. Logging events are divided into five classes:

CLAS0010 records internal errors.

CLAS0011 records APPC subsystem warnings.

CLAS0012 records APPC subsystem information messages.

CLAS0013 records conversation information messages.

CLAS0014 records session performance statistics.

You can configure the logging messages to go to the system console, to the log file, and to three separate user.accounts. Critical messages should be logged to both the system console and the log file. That way, messages erased from the terminal buffer can still be found in the log file. Hewlett-Packard recommends that all classes of logging messages be recorded in the log file and that all classes except CLAS0014 (performance statistics) be logged to the system console.

The logging messages and their meanings are documented under "Logging Messages," in Appendix A “Messages” of this manual. For more information on logging and logging configuration, see the SNA Link/XL Node Manager's Guide.

APPC Subsystem Messages and Return Codes

Appendix A “Messages” contains a complete list of all APPC subsystem messages and return codes. It describes the cause of each message and any actions you should take to resolve a problem. The APPC subsystem generates the following kinds of messages:

The SNACONTROL STATUS Command

To get the status of the local SNA node and LU, issue the SNACONTROL STATUS command. This command will tell you whether SNA Transport is active and will display the status of the PU-SSCP sessions and LU-SSCP sessions.

	NOTE: For Type 2.1 nodes, the SNACONTROL STATUS command always lists the SESSION STATE as RESET (unless Node 2.1 Dependent LU Support = Y in the "SNA Node Configuration: PU Data" screen). To find out whether a Type 2.1 node is active, issue the command SNACONTROL STATUS;LINK=LINKSTATE The SESSION STATE will still be listed as RESET, but in the upper right corner of the display, the Linkstate will tell you whether the node has been activated. The Linkstate should be CONNECTED LEVEL 2.

See the SNA Link/XL Node Manager's Guide for a complete description of the SNACONTROL STATUS command.