1

I am following the guide for capturing a trace in SQL Server Database Experimentation Assistant here: https://docs.microsoft.com/en-us/sql/dea/database-experimentation-assistant-capture-trace?view=sql-server-2017

I am using DEA version 2.6. The article above states that I should be able to choose the capture trace file size for rolling over files. However, in version 2.6 I do not have the option to set trace file size.

As a result, I end up with dozens of 500 MB files at the end of the capture. Then I am expected to sit there and replay each trace file, one by one. Is there any easier way of doing this, or of forcing the trace file to grow larger so I end up with less rollover files?

My new capture screen: enter image description here

MS doc capture screen: enter image description here

1

I end up with dozens of 500 MB files at the end of the capture.

Yes I agree the documentation is wrong and you cannot configure file size. You can only configure duration. 500 MB is by design and recommended here and here. 200 MB or 512 MB is the ideal trace rollover size to use to capture production or performance traces. Second link give you few reasons why this is the recommended size. One of them is:

  • If you use our "published" versions of production and performance trace definition files and the performance trace definition filters out the traces for one database of interest, analysis shows that the Performance Trace size is about 15 times larger than the Production Trace size.

Then I am expected to sit there and replay each trace file, one by one. Is there any easier way of doing this, or of forcing the trace file to grow larger so I end up with less rollover files?

You do not need to process each file individually. In Full path to source File instead of individual file, give the folder name where your raw trace files are located. Tool will create a new sub-folder named PreProcessOutput and put all Replay Markup Language (RML) files there.

enter image description here

On a side note hopefully you configured your Distributed Replay successfully.

  • Thank you!! I will try your tip to just specify the folder, rather than an individual trace file. I didn't even bother trying to set up distributed replay....I'm just going to go with one server replaying; concurrency is not worth the hassle of setting up the DR agents and controller. – blizz Jun 22 at 2:19
  • Just a note - for some reason I had to run DEA as Administrator in order to get this working – blizz Jul 3 at 15:05

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.