2 Replies Latest reply on Jun 3, 2019 9:19 AM by Feli Brachthaeuser

    MVDB2 - PR: DPDSPIN - how many spin datasets are Best Practice?

    Feli Brachthaeuser
      Share This:


      MVDB2 Performance Reporter has DPDSPIN Datasets which are docuemnted by this:


      Defines a spin file to retain SMF statistics records that end a period to resolve intervals on subsequent SMF processing

      The spin file is useful only when the same input source (such as an SMF Generation Data Group or Data Collector archive) is used for each DPRSMF run. When the DPRSMF job runs, information from the last SMF statistics record encountered for each DB2 subsystem is written to the spin file and then used as the first record for the next DPRSMF run. The DPRSMF job must run at regular intervals to ensure that no intervals are lost.

      If DPRSMF does not run at regular intervals, allocate a new spin file for each run. DMRSMF allocates the spin file with DISP=OLD. To create a new spin file, use the BBSAMP sample JCL member DPJSPIN.

      Step 2—Statistics consolidation - Documentation for MainView for DB2 12.1 - BMC Documentation

      As my customer is running various DPRSMF using the archives from the data collector as input, we are wondering how many DPDSPIN datasets we will need.

      Up until now, there has only be one, which is more related to the RTE than the environments as test and prod.

      It can easly happen that multiple jobs are running concurrently and more than one time per day.

      And DPDSPIN is allocated with DISP=OLD.

      What is your recommendations to prevent locking and data loss?