Updating init ora
Each instance can have more than two redo log groups, but two are minimum per instance.Each instance is assigned a thread number starting at 1, when you add more logfiles make sure you mention the THREAD, so that logfiles will be added to right instance.In part 1 I explained how to duplicate a single instance database using RMAN.This article will focus on converting the single instance to RAC. Convert Single instance to RAC We have migrated the database successfully from Non-ASM to ASM using RMAN; the database is still running as single instance.Create second thread of online redo logs in order to start instance 2 and enable thread 2.It is always recommended to create redo log groups across disk groups for redundancy: If you have more than two nodes then you need to repeat the steps on each node in the cluster.Generally speaking, redo and undo are handled on a per instance basis.You must create two redo log groups for each instance and all redo groups must be stored on shared devices for an instance or crash recovery purpose.
Hi experts, After I upgraded my system Oracle 11g to Oracle 12c I have the error ORA-04031 every week, Errors in file I:\ORACLE\PRO\SAPTRACE\diag\rdbms\pro\pro\trace\pro_arc2_8028(incident=530484): ORA-04031: unable to allocate 352 bytes of shared memory ("shared pool","unknown object","sga heap(3,0)","krsdicle")Errors in file I:\ORACLE\PRO\SAPTRACE\diag\rdbms\pro\pro\trace\pro_arc0_14736(incident=514716): ORA-04031: unable to allocate bytes of shared memory ("","","","")I have to stop the database and started again...
I increase the parameter shared_pool_size but only work one-two days more... I have a server with 64GB and the parameters are:_enable_shared_pool_durations boolean FALSEshared_pool_reserved_size big integer 405600Kshared_pool_size big integer 5Gsga_max_size big integer 32Gsga_target big integer 0memory_target big integer 0pga_aggregate_limit big integer 12Gpga_aggregate_target big integer 6GAny idea?
Now we need to convert the single instance database to RAC.
2.1 Create redo and undo for second instance Each individual instance requires an independent set of redo and undo segments.
2.2 Add cluster related parameters The duplicated instance does not have any cluster related parameters, we need to add the cluster parameter in order to convert single instance to RAC.
The CLUSTER_DATABASE=true parameter needs to set before a database can be started in cluster mode.