Steps to clone database in other server with restore daily

Question

Hi,
I think i know what is the steps but need a guru that give me the ok.
I have been asked to perform a clone database (11.2.0.3 with ASM) in another server. The original database (call PROD) has full backup everyday that it automatically (with crontab and a script i guess) restore the second database. The problem i see its that the original database is a RAC (2 nodes) but the second database is standalone (call TEST)
I think the steps are:
1) create a standalone database in second server with same number of diskgroups and same directories
2) copy full backup to TEST server.
3) duplicate with rman:
connect auxiliary sys/password@TEST
connect target /
startup auxiliary nomount;
run {
allocate auxiliary channel ch1 type disk;
duplicate target database to 'TEST' nofilenamecheck;
release channel ch1;
4) copy PFILE from PROD and modify (like standalone database) at TEST
5) startup nomount; --test database
6) restore controlfile from '< backup piece name of controlfile auto backup copied in step 2>'; --at test database
7) alter database mount; --test database
8) recover, restore from full backup and open database-->
run {
allocate channel d1 device type disk format ' <path of the copied backup in step 2 >';
set until sequence < sequence> thread < thread# >;
restore database;
recover database;
SQL> alter database open resetlogs;
9) like its a backup that has 2 undo tablespaces, should i drop 1?:
SQL> drop tablespace UNDOTBS2 including contents and datafiles;
All this steps at first time, are they right?, to do it automatically in the script i should use 2) 5) 6) 7) 8)?
-- like the databases names are different, should i have any problem when restore de controlfile? if yes, how to avoid it?
Thanks in advance and sorry for the inconveniences
Edited by: Me_101 on 30-jul-2012 5:46

Answer 1

Me_101 wrote:
Hi,
I think i know what is the steps but need a guru that give me the ok.
I have been asked to perform a clone database (11.2.0.3 with ASM) in another server. The original database (call PROD) has full backup everyday that it automatically (with crontab and a script i guess)Why guess? Can't you look at the crontab schedule and see for yourself?
restore the second database. The problem i see its that the original database is a RAC (2 nodes) but the second database is standalone (call TEST)
That should not be an issue. Even with RAC there is still only a single database to duplicate.
I think the steps are:
1) create a standalone database in second server with same number of diskgroups and same directories
2) copy full backup to TEST server.
3) duplicate with rman:
connect auxiliary sys/password@TEST
connect target /
startup auxiliary nomount;
run {
allocate auxiliary channel ch1 type disk;
duplicate target database to 'TEST' nofilenamecheck;
release channel ch1;
4) copy PFILE from PROD and modify (like standalone database) at TEST
5) startup nomount; --test database
6) restore controlfile from '< backup piece name of controlfile auto backup copied in step 2>'; --at test database
7) alter database mount; --test database
8) recover, restore from full backup and open database-->
run {
allocate channel d1 device type disk format ' <path of the copied backup in step 2 >';
set until sequence < sequence> thread < thread# >;
restore database;
recover database;
SQL> alter database open resetlogs;
9) like its a backup that has 2 undo tablespaces, should i drop 1?:
SQL> drop tablespace UNDOTBS2 including contents and datafiles;
All this steps at first time, are they right?, to do it automatically in the script i should use 2) 5) 6) 7) 8)?
-- like the databases names are different, should i have any problem when restore de controlfile? if yes, how to avoid it?
Thanks in advance and sorry for the inconveniences
Edited by: Me_101 on 30-jul-2012 5:46Take a look at the DUPLICATE DATABASE command in rman. Key is to connect to the production db as target, duplicate databse as auxiliary
connect catalog rman/cat@rmcat
connect target sys/pswd@proddb
connect auxiliary /
run {                        
duplicate target database to NEWDBSID;
}                                        where NEWDBSID is the name of the new database
Put together what you think is the correct procedure and give it a shot. If you have specific problems, bring them to the forum.