4 Replies Latest reply on Jan 23, 2014 10:36 AM by user11240176

    Recover in standby in cluster

    975791

      Hi,

       

      We have Primary & standby in cluster.All are 3 node.

      Could we start recovery in all standby instances? or do i need to start only from one node?

       

      When executed from one of the node,we got the below outout.

       

      SQL> ALTER DATABASE RECOVER MANAGED STANDBY DATABASE USING CURRENT LOGFILE DISCONNECT;
      ALTER DATABASE RECOVER MANAGED STANDBY DATABASE USING CURRENT LOGFILE DISCONNECT
      *
      ERROR at line 1:
      ORA-01153: an incompatible media recovery is active


      SQL> !
      [oracle@plm-20 /u01/app/oracle/product/11.2.0.3/db_1/dbs]$ ps -ef| grep mrp
      oracle   10620  9869  0 13:13 pts/0    00:00:00 grep mrp

       

      Br,

      Raj

        • 1. Re: Recover in standby in cluster
          jgarry

          You can usually get a clue when you say ALTER DATABASE rather than ALTER SYSTEM.  The -i parameter on grep is sometimes useful.

          • 2. Re: Recover in standby in cluster
            Baris Yildirim

            Hi,

            Did you try again?

             

            is there a running rman backup?

             

            Regards

            • 3. Re: Recover in standby in cluster
              Alvaro

              You should be able to start the media recovery on any standby node. Just check your SERVICE_NAME on the primary that you use on the log_archive_dest also accounts for the new instance of the standby.

               

              Referring to your problem, make sure you stopped the MRP on the prior node before trying to start it on the new node.

               

              Only one node can be running the MRP0, writing to the standby database. The others can be mounted (or even opened read only in case of active dataguard configuration - license) but not running the managed recovery process.

              • 4. Re: Recover in standby in cluster
                user11240176

                As already explained by Alvaro, recovery needs to run only one node on a clustered standby and rest all nodes can have instance open read only in case of ActiveDG.

                Please ensure if recovery process is already active on other nodes. For eg:

                 

                SQL>  select INST_ID,PROCESS,STATUS from GV$MANAGED_STANDBY where PROCESS like 'MRP%';

                   INST_ID PROCESS                              STATUS
                ---------- ------------------------------------ ------------------------------------------------
                         1 MRP0                                 APPLYING_LOG

                 

                Thanks,

                Asif Haliyal