Understanding Tracing

Tracing is done in the code of IBM Security Directory Integrator using JLOG's PDLogger object. PDLogger or the Problem Determination Logger logs messages in Logxml format (a Tivoli® standard), which IBM® Support understands and for which they have processing tools.

The basic level of information traced, as handled by the PDLogger APIs, is:

Date | Time | ClassName | methodName | MachineName | IP | {Entry/Exit/Exception} | [Parameter]

Basic tracing information means Time, Level (Min, Mid, Max), Location in code, that is Method name and Entry/Exit. The "|" character serves a documentation purpose only, it is not part of the actual log.

Tracing is not performed using Log4J Appenders for the following reasons:

Trace is always to be enabled
You wouldn't want multiple traces enabled in the server (could be several for each AL if Appenders were used).

The PDLogger is attached to the JLOG SnapMemory handler and the JlogSnapHandler.

The SnapMemory Handler logs trace messages to memory. On the trigger of a LogEvent (that is, an occurrence of a specific Log level Trace message, as defined by the jlog.levelflt.level filter, or an application crash or on the occurrence of a specific TMS XML messageID) the Trace memory buffer is written to a file by the JlogSnapHandler.

To make Tracing and Log messages in IBM Security Directory Integrator unique across all IBM products, they are prefixed with a unique prefix: CTGDI.

All error messages are prefixed with a unique TMSXML messageID that indicates the cause of the error and an operator response.

All info messages are also prefixed with a unique TMSXML messageID that may or may not provide the operator response.