Trace TCP applications

You can use the quality of service (QoS) trace to work with trace functions and to see the current trace buffer.

To run the trace on the system, type TRCTCPAPP (Trace TCP/IP Application command) from a command-line interface.

The following table introduces the possible parameters to use in a trace. If a setting is not displayed on the character-based interface, enter it in a command. For example, TRCTCPAPP APP(*QOS) MAXSTG(1000) TRCFULL(*STOPTRC) ARGLIST('l=4 c=i').

Settings	Options
TCP/IP application	QOS
Trace option setting	ON, OFF, END, CHK
Maximum storage for trace (MAXSTG)	1-16000, *APP
Trace full action (TRCFULL)	WRAP, STOPTRC
Argument lists (ARGLIST)	Levels: 'lvl=1', 'lvl=2', 'lvl=3', 'lvl=4' Content: 'c=a', 'c=i', 'c=d' , 'c=m'
QoS trace type	*ALL

Maximum storage for trace

1-16000: This is the maximum storage size for the trace data. The trace either stops or wraps when this size is reached. The default size is 4 MB. To specify the default size, select *APP.
*APP: This is the default option. It tells the application to use its default trace size. The default trace size for the QoS server is 4 MB.

Trace full action

*WRAP: Wraps the trace information when the trace reaches the maximum disk space (trace buffer size). Wrapping allows the system to overwrite the oldest information in the file so that you can continue recording the trace information. If you do not select wrap, then the trace operation stops when the disk is full.
*STOPTRC: Stops collecting information when the system reaches the maximum disk space.

Argument lists

The argument lists specify which error levels and content are logged. There are two arguments allowed in the TRCTCPAPP command: trace level and trace content. When you specify the trace level and trace content, make sure all attributes are contained in a single set of quotation marks, for example, TRCTCPAPP 'l=4 c=a'

Log levels are inclusive. This means that when you select a log level, all previous log levels are also selected. For example, if you select level 3, then levels 1 and 2 are automatically included. In a typical trace, it is recommended you specify 'l=4'.

Trace levels

Level 1: System errors (SYSERR): Logs errors that occur in systems operations. If this error occurs, the QoS server cannot continue. For example, a system error might occur if you are running out of system memory or if your system cannot communicate with TCP/IP. This is the default level.
Level 2: Errors between objects (OBJERR): Logs errors that occur within the QoS server code. For example, an object error might occur because a system operation encounters some unexpected result. This is generally a serious condition that must be reported to service.
Level 3: Specific Events (EVENT): Logs any QoS operation that has occurred. For example, an event log records commands and requests. The results are similar to the QoS journaling function.
Level 4: Trace messages (TRACE): Traces all data being transferred to and from the QoS server. For example, you might use this high-level trace for logging anything that you think might be helpful for debugging problems. This information is helpful to determine where a problem occurred and how to reproduce the problem.

Trace content

Only specify one content type. If you do not specify what content to trace, then (by default) all content will be traced.

Content = All ('c=a'): Traces all functions of the QoS server. This is the default value.
Content = Intserv ('c=i'): Traces the integrated service operations only. Use this if you determine the problem to be integrated service related.
Content = Diffserv ('c=d'): Traces the differentiated service operations only. Use this if you determine the problem to be differentiated service related.
Content = Monitor ('c=m'): Traces the monitor operations only.

If you need help interpreting the trace output, read the trace output example on the trace output page, which contains sample output with comments to help you interpret its meaning. The TRCTCPAPP function is typically used by the service, so if you have problems reading the output, you might contact your service representative.

Examples: Reading the trace output
This is not an all-inclusive discussion of how to read your trace output. However, it does highlight the key events to look for in the trace information.

Parent topic:

Troubleshooting QoS

Related reference
Trace TCP/IP Application (TRCTCPAPP)