2 Replies Latest reply on Apr 11, 2014 1:51 PM by mseberg

    Failover test *without* 'fast-start failover' config: how to reboot primary clean and convert it to standby ?




      We're currently trying to do some tests with our 2 primary & standby RAC databases in 11gR2.

      1/ we -still ?- don't use the "fast-start failover" observer config;


      2/ we are going to power-off both nodes on the primary (to simulate a loss of this site),

      without modifying the Dataguard config;


      3/ on the standby site I'll convert my DB to primary role (I know how to do that:

      SQL> alter system set dg_broker_start=false scope=both sid='*';

      SQL> alter database recover managed standby database finish;

      SQL> alter database commit to switchover to physical primary;


      4/ then we will restart the primary site nodes - but this is here that I have a problem:

      Oracle is going to start up everything, including Dataguard, and I know it is going

      to be a serious issue (cause we already experienced it): how to see to it that this

      primary DB once restarted will change its role to standby ? Should I then select

      the STANDBY_BECAME_PRIMARY_SCN from V$DATABASE (in my newly primary

      on previous standby site) and flashback my ex-primary to this scn as quick as I can

      once the nodes are up again ? Will  that be enough ?


      To sum up anybody ever experiment that kind of d-i-y failovers without setting

      up fast-start failover ?


      Thanks a lot in advance...

      Kind regards,