seamless physical dataguard creation and swichover

prerequest : primary database name is ocmdb , standby database's directory strutrue is different than primary (ocmdb,ocmdg) maximize protection mode

1. create database

2. alter database force logging

3. orapwd

4.modify primary initocmdb.ora setting following parameters: primary db_unique_name=ocmdb standby db_unique_name=ocmdg

fal_server=ocmdg

fal_client=ocmdb

log_archive_dest_1='location=/u01/oracle/oradata/arch/ocmdb valid_for(all_logfiles,all_roles) db_unique_name=ocmdb'

log_archive_dest_2='service=ocmdg lgwr sync affirm valid_for(online_logfiles, primary_role) db_unique_name=ocmdg'

db_file_name_convert=(ocmdb,ocmdg)

log_archive_config='dg_config=(ocmdb,ocmdg)'

standby_file_management=auto

user_dump_dest=/u01/oracle/admin/ocmdb/udump

core_dump_dest=/u01/oracle/admin/ocmdb/cdump

backgroud_dump_dest=/u01/oracle/admin/ocmdb/bdump

db_unique_name=ocmdb  -- this parmater can garantee the database has a service name that be registered by listener

5. on primary : alter database create standby controlfile as '/u01/control.ctf' ;

6.shutdown primary and scp /u01/oracle/oradata/ocmdb/* oracle@db2:/u01/oracle/oradata/ocmdg  scp $ORACLE_HOME/dbs/initocmdb.init oracle@db2:/$ORACLE_HOME/dbs and modify initocmdb.ora on standby and replace the standby control file with standby control file. orapwd on standby database

 

fal_server=ocmdb

fal_client=ocmdg

log_archive_dest_1='location=/u01/oracle/oradata/arch/ocmdb valid_for(all_logfiles,all_roles) db_unique_name=ocmdb'

log_archive_dest_2='service=ocmdg lgwr sync affirm valid_for(online_logfiles, primary_role) db_unique_name=ocmdg'

db_file_name_convert=(ocmdb,ocmdg)

log_archive_config='dg_config=(ocmdb,ocmdg)'

standby_file_management=auto

user_dump_dest=/u01/oracle/admin/ocmdg/udump

core_dump_dest=/u01/oracle/admin/ocmdg/cdump

backgroud_dump_dest=/u01/oracle/admin/ocmdg/bdump

db_unique_name=ocmdg

 

7.startup mount primary

8. create standby logfile on primary , attention is standby logfile should be same size to normal logfile ,otherwise max protection will fail.

alter database create standby logfile group 4('/u01/oracle/oradata/ocmdb/dredo04.log') size 5m;

alter database create standby logfile group 5('/u01/oracle/oradata/ocmdb/dredo05.log') size 5m;

alter database create standby logfile group 6('/u01/oracle/oradata/ocmdb/dredo06.log') size 5m;

alter database create standby logfile group 7('/u01/oracle/oradata/ocmdb/dredo07.log') size 5m;

9.startup mount on standby database , make the primary and standby database both at least at mount state , this way pmon can auto find and register the database on listener

10. netca to configue the listener and tnsnames.ora all using normal mode on both primary and standby  , lsnrctl start and lsnrctl status make sure all the tns alias work fine.

11. create standby logfile file group on standby database

12.alter the standby database in recover mode : alter database recover managed standby database using current logfile disconnect;

13.on primary alter database open , check select error from v$archive_dest , if no error ,issue the command alter database set standby database to maximize protection;

14.alter database commit to switchover to [primary|physical standby];

15.open a terminal both on primary and standby and tail -f alert log to check the state of database