Log files

If you have Administrator authorization, you can determine the characteristics of the log files associated with each output group in the log set. You can define:

• Required time span of data that you want immediately available to users in days, hours or minutes.
• Minimum and maximum number of data sets along with their location

At product initialization the minimum number log files are allocated. As the space in these data sets is filled, new log files are created until either the time span is reached or the maximum number of data sets is allocated.

When all of the allocated log files are full, the oldest log data set is reused and existing data is over-written.

The log set definition also includes the option to copy trace data from an active log file by using a started task with the default name NGLARCH. The product activates NGLARCH when the log file is full. The archive copy of the data is written and information about the archive file is recorded in the Archive Directory (also known as the COPYDIR).

To assure that no data is lost during the transition from one log file to another and the copy of the older data to an archive file there should always be a minimum of at least two log files for each log set.

The following figure shows the flow of data from Db2 to the trace data sets and shows a sample of how the output groups for a DOMPLEX Profile might be configured.

1. Interval data is written to a table in storage and summarized until at the end of the defined interval.
2. The object data is written to the log files associated with Output Group 004 and the accounting data is written to the log files associated with Output Group 003.
3. Event data from the three Db2 subsystems is written immediately to log files. SQL text, host variables and exception data is written to the log files associated with Output Group 002.