Application Development Discussions
Join the discussions or start your own on all things application development, including tools and APIs, programming models, and keeping your skills sharp.
cancel
Showing results for 
Search instead for 
Did you mean: 

trace file is rolling over

Former Member
0 Kudos

hi experts,

I am tracing an ABAP application using ST05. But the program is so large and the records are too many, that the trace records are getting rolled over and I am able to see only the last set of trace records. The first group is getting rolled over. Is there any suggestion to avoid this and thus enabling me to view all the records. Thanks in advance,

anirvesh,

3 REPLIES 3

Former Member
0 Kudos

If there is, I don't know it.

But what I would do is to execute the program in chunks and turn the trace on and off for each chunk.

Or you could use the runtime analysis (SE30) to identify the problem area and just do ST05 on that.

Rob

Former Member
0 Kudos

Hi Anivesh,

the parameters you're looking for are <i>rstr/max_filesize_MB</i> and <i>rstr/max_files</i>. We have them set to 64 and 20 (default: 16 and 10 afaik). This means, that we could store traces up to 2064MB = 1280MB (default 1016MB = 160MB). You have to have the disk space on your application server because the system trace writes to the instance log directory.

With our settings we are even able to trace huge CO transactions. But you should keep in mind, that a huge trace needs quite some time to be analyzed by ST05. So you need to be patient and you have to have enough memory available to display the trace file content.

Regards

Ralph

former_member194613
Active Contributor
0 Kudos

The problem with overwritten SQL traces is rare in newer SAP releases, maybe you should check the parameter settings.

Very large traces are hard to analyze, because I would absolutely recommend you the start you analysis with SQL statement summary, see detail here:

SQL trace:

/people/siegfried.boes/blog/2007/09/05/the-sql-trace-st05-150-quick-and-easy

With the SQL Trace it no problem, to stop the program at breakpoint and stop the first measurement. Larger continue and start the second measurement.

Never forget, to do preruns to fill buffers and caches, because you can only optimize the repeated executions.

Siegfried