3 Replies Latest reply: Jul 29, 2013 7:26 AM by mseberg RSS

    Parallel recovery on DR on RAC

    1010036


      Hi all,

       

      We have a 2 node DR on 2 node RAC. Just wanted to know if it is possible to do recovery from both the nodes parallel, I mean node 1 will apply its archive and node 2 prod nodes 2 archive.

      If this is possible how is that and if its not why?

       

      Regards

      ID

        • 1. Re: Parallel recovery on DR on RAC
          mseberg

          Hello;

           

          Have you looked at:

           

          D.1.1 Setting Up a Multi-Instance Primary with a Single-Instance Standby

           

          Data Guard Concepts and Administration 11g Release 2 (11.2) E25608-04

           

          If yes, what would parallel recovery give you?

           

           

          I have never tried to do this setup but the main issues I would see it the second standby instance would not be available

          if something happened to the first standby instance.

           

          By using only one instance for apply it's much simplier to keep track of the redo coming from the two primary instances.

           

           

           

          If you don't use a standby as a reader database it does not have a lot of load on it. Generally if you check the alert

          log there are several minutes between the "waiting for log..." messages.

           

          Great question!

           

          Best Regards

           

          mseberg

          • 2. Re: Parallel recovery on DR on RAC
            1010036

            Thanks for the reply,

             

            I am still not conveinced, i believe there should be some reason behind it.

             

            Regards

            ID

            • 3. Re: Parallel recovery on DR on RAC
              mseberg

              Hello again;

               

               

              Do you have redo generation so high that one server cannot keep up? If yes I would look at setting LOG_ARCHIVE_MAX_PROCESSES higher.

               

              Compression is another option. You can also Configure the TCP "Send" and "Receive" Buffer Sizes (RECV_BUF_SIZE / SEND_BUF_SIZE)

               

              Check you SDU Size and increase if needed.

               

              If you want an exact reason for you first question I think somebody at Oracle will have to answer that.

               

              Best Regards

               

              mseberg