2 Replies Latest reply on Jul 22, 2021 6:32 AM by Kenneth Guyette

    Incremental Profiling for MS SQL Server

    Kenneth Guyette Seasoned Veteran

      I am trying to get a better understanding of how Incremental Profiling works in EDC.  I know conceptually that it should only grab things that have changed since the last profiling run but I am not clear on how this works systematically.   For example, I am trying to understand if it is comparing the last profiling date against the system date on each table and/or view (for MS SQL Server) to determine which ones to profile again?  I am trying to gauge what type of performance hit this is on the job.  In theory we would be profiling less but we would have to check each table/view to see what would be profiled.  I assume the performance hit would be on the EDC side on not on the server being profiled.

       

      While researching this I came across this....

      EIC-52725 : Incremental profiling is not working for MS SQL Server in EDC 10.5

      How can I find more information on what this means?  The title is pretty descriptive but is it not working at all???

        • 1. Re: Incremental Profiling for MS SQL Server
          Shruthi Sadashiva New Member

          Here is more detail on the current issue

           

          1.Run the profiling on SQL server without enabling Incremental profiling. 23 tables present in DB and 23 tables sent for profiling

           

           

          2.The second run added 2 more tables and selected Incremental Profiling

           

          Now ideally the expectation is to have 2 tables sent for profiling. But we see 25 tables sent for profiling.

          We see 25 logs created in $INFA_HOME/ logs/node01/services/DataIntegrationService/disLogs/profiling which confirms incremental profiling is not working

          Not honoring Support

          • 2. Re: Incremental Profiling for MS SQL Server
            Kenneth Guyette Seasoned Veteran

            Thanks for the explanation.  Assuming this were working and going back to my original question, how would it identify the two that needed to be scanned.  Just trying to understand where the load would be in terms of performance as our DBAs already have some concerns about the performance hit on the source SQL servers with the initial loads.  We were hoping that incremental profiling would alleviate some of that concern.

            This not working seems to be a pretty big issue.  Is there any type of ETA on when this will be addressed?