This content has been marked as final. Show 10 replies
I'm doing physical dataguard configuration.
The primary database is running perfectly
the secondary database is created from the primary database, and In order to custom the init parameter of the secondary database, I need to create a pfile from the spfile and them customize the secondary's database pfile.
That's why I use the pfile just when I created the secondary instance , after that I create the spfile from the pfile.
Every thing works smoothly but I'm wondering why the instance doesn't startup when the OS is started?, It's because this instance is a standby database?
Standby databases must be started manually, however you maybe to script the process. I'm currently implementing a standby database, so I'll have to check that out myself. You probably have an error in your alert log about can't start up database from standby control file, or something to that affect.
Oracle recommends you use spfile with data guard. I'm using 10g Grid R3 and used it to create the standby database. Don't have to customize anything, it does it all for me. The one thing you do have to do, is the switchover management. TAF is not designed for switchover, its designed for fail over. I created a second listener for 10G Grid and the Database broker to communicate between the primary and standby. If I need to switch over, I just click a button in 10g Grid, then shutdown down the listener used by the application.
FYI, 10G Grid R2 has issues with data guard.
I just used the pfile at the first time I created the instance, and after that if I found that every thins is OK, I created the spfile from the pfile.
The issue is only when I shutdown and startup the whole system, the secondary database doesn't startup, and I have to do that manually, this is my issue.
I don't have 100% control of the secondary server that is why I don't need to give the system admin how to start the database manually.
I also know that eventhough there is a gap if the secondary database is automatically the FAL server can resolve automatically the gap.
also wondering about GRID CONTROL? should I need extra license
Maybe I didn't explain exactly what I was expecting with the standby database, these steps are what I did, to resolve the issue
ENABLE THE STANDBY DATABASE TO STARTUP FROM SPFILE
Step 1: create a newly spfile from the current pfile
SQL> CREATE SPFILE='C:\ORACLE\PRODUCT\10.2.0\DB_1\DATABASE\SPFILEORCL.ORA' FROM PFILE='C:\ORACLE\PRODUCT\10.2.0\DB_1\DATABASE\INITORCL.ORA';
Step 2: Rename the INITORCL.ORA TO SAVE_DATE_INITORCL.ORA
Example: the C:\ORACLE\PRODUCT\10.2.0\DB_1\DATABASE\INITORCL.ORA became
Step 3: delete the former sid created
C:\ORADIM -delete -sid orcl
Step 3: Change the Oracle service to start when the OS start
C:\ ORADIM -new -sid ORCL -SRVC OracleServiceORCL -STARTMODE auto -SRVCSTART system -SPFILE
Now When I shutdown the whole system and start it up, the standby database startup. the only thing is that the standby database startup a read-only mode, defautl with oracle 10g this is fine for me, because the MRP is stopped but the log shipping is not stopped.
So eventhough the system admin shutdown the database server at OS level without letting me know, tha standby should continue receiving archived log files.
I'll just need to monitore the lop apply process and start it to apply archived log files received
But once the database is started in read-only mode, don't you have to shut it down to put it back in recovery mode to start the log apply process? I'm not sure why you want to do that.
If you script it, you won't have to worry about it, say if you are on vacation......unless you are worried about job security.
This is the old fashion but also work with 10g
ALTER DATABASE MOUNT STANDBY DATABASE;
ALTER DATABASE RECOVER MANAGED STANDBY DATABASE DISCONNECT FROM SESSION;
Default mode with oracle 10g when starting up the dataguard
STARTUP; ---> read only mode by dafault
ALTER DATABASE RECOVER MANAGED STANDBY DATABASE DISCONNECT FROM SESSION; --> recovery managed mode, means the secondary database is recovered from the archived log files
ALTER DATABASE RECOVER MANAGED STANDBY DATABASE USING CURRENT LOGFILE DISCONNECT FROM SESSION; -->managed Real Time Apply means the secondary database is recovery from the redo data fron the standby redo log files as soon as they are filled up by the RFS.
That's why sometimes using tools hide some details, like Grid Control, this is the reason why I prefer first to catchup all step manually and them go further with the grid control.
Any way Job secure is not my concern only be sure to strongly explain to my boss why choosing this direction than other.