3 Replies Latest reply on Mar 28, 2013 1:03 PM by mseberg

    Does thread# matter for standby redo logs in a RAC-to-RAC DataGuard

      Oracle RAC using DataGuard
      Primary 3 nodes
      Standby 2 nodes

      We created our standby redo logs (SRLs) as thread 1. There are 2 nodes for the standby instance. Since the standby only works from one apply instance, do the SRLs need to be threaded?

      From Oracle documentation:
      The THREAD clause is required only if you want to add one or more standby redo log file groups to a specific primary database thread. If you do not include the THREAD clause and the configuration uses Real Application Clusters (RAC), Data Guard will automatically assign standby redo log file groups to threads at runtime as they are needed by the various RAC instances.

      This statement is confusing because it references threads in relation to the primary, but SRLs are used only on the standby.

      Is there a best practice to define threads for each instance, or will 1 thread do just as well? Performance reasons, perhaps?

      We added SRLs to the standby without defining thread, and we immediately saw it assigned 0. The next morning we see it is assigned 2. Is this a good thing, a better way of managing the SRLs?

      Also, we thought that standby_file_managment=AUTO would allow creation of SRL at primary (to prepare for role reversal), and it would be propagated to the standby, but that didn't happen.
      On the standby, we had to cancel the log apply and set standby_file_management=MANUAL, add the SRLs, and then set back to AUTO and start the apply process.

      Any insights into SRL management in a DG environment would be appreciated.