Another RMAN duplicate problem - RAC database to single instance

Question

Hi,
I have a problem with the RMAN duplicate procedure and was hoping someone can help.
I would like to create a duplicate of our production RAC database on a separate, stand-alone, database server on another site. This duplicate will be used for intensive querying by another business unit who I don't want to have access to our production database.
My procedure goes like this:
1. Create a disk (not ASM) based backup of the datafiles and any archived redo logs:
"run {
allocate channel d1 type disk;
backup format '/u02/stage/df_t%t_s%s_p%p' database plus archivelog delete input;
release channel d1;
2. Tar and scp these files to the same location on the stand-alone database server.
3. In the meantime, work has been happening on the production database and further archived redo logs have been generated. I don't really care about these logs for the purposes of the duplicate however, I just want to duplicate to the point of the recent backup. To do this, I run the following SQL to determine the sequence number that I should be duplicating up to:
"select max(sequence#), thread# from v$archived_log where deleted='YES' group by thread#;"
4. Duplicate the production database to the stand-alone database (all the SQL Net stuff is working).
"run {
set until sequence (value returned by above SQL statement);
duplicate target database to XXX;
However, my problem arises because I don't know how to handle the fact that there are two threads. I understand that each thread relates to one of the RAC instances, I just don't know which one to specify for the duplicate. We have a database service which the client application connects through, and that service runs on on or other of the instances. Should I just care about the logs from the instance where the service is currently running?
Am I even approaching this is the correct way?
I look forward to any help that people may be able to offer.
Regards,
Phil

Answer 1

Hi Werner,
Thanks again for your help, there is still something wrong though. "list backup of archivelog all;" shows:
BS Key Size Device Type Elapsed Time Completion Time
3784 202.34M DISK 00:00:08 28-OCT-09
BP Key: 3784 Status: AVAILABLE Compressed: NO Tag: TAG20091028T111718
Piece Name: /u02/stage/df_t701435838_s3820_p1
List of Archived Logs in backup set 3784
Thrd Seq Low SCN Low Time Next SCN Next Time
1 9746 569095777 28-OCT-09 569150229 28-OCT-09
1 9747 569150229 28-OCT-09 569187892 28-OCT-09
1 9748 569187892 28-OCT-09 569231956 28-OCT-09
1 9749 569231956 28-OCT-09 569259816 28-OCT-09
2 7931 569095774 28-OCT-09 569187902 28-OCT-09
2 7932 569187902 28-OCT-09 569259814 28-OCT-09
BS Key Size Device Type Elapsed Time Completion Time
3787 1.04M DISK 00:00:02 28-OCT-09
BP Key: 3787 Status: AVAILABLE Compressed: NO Tag: TAG20091028T112222
Piece Name: /u02/stage/df_t701436142_s3823_p1
List of Archived Logs in backup set 3787
Thrd Seq Low SCN Low Time Next SCN Next Time
1 9750 569259816 28-OCT-09 569261110 28-OCT-09
2 7933 569259814 28-OCT-09 569261108 28-OCT-09
You can see that the highest sequence number is 9750 of thread 1, and that the Low and Next SCNs are 569259816 and 56926111. However, when I look at the output of the RMAN duplicate command:
contents of Memory Script:
set until scn 569505448;
recover
clone database
delete archivelog
executing Memory Script
executing command: SET until clause
Starting recover at 28-OCT-09
allocated channel: ORA_AUX_DISK_1
channel ORA_AUX_DISK_1: sid=39 devtype=DISK
starting media recovery
Oracle Error:
ORA-01547: warning: RECOVER succeeded but OPEN RESETLOGS would get error below
ORA-01194: file 4 needs more recovery to be consistent
ORA-01110: data file 4: '/u02/sca-standby/data/users.260.623418479'
RMAN-03002: failure of Duplicate Db command at 10/28/2009 16:12:55
RMAN-03015: error occurred in stored script Memory Script
RMAN-06053: unable to perform media recovery because of missing log
RMAN-06025: no backup of log thread 2 seq 7936 lowscn 569411744 found to restore
RMAN-06025: no backup of log thread 2 seq 7935 lowscn 569321987 found to restore
RMAN-06025: no backup of log thread 2 seq 7934 lowscn 569261108 found to restore
RMAN-06025: no backup of log thread 1 seq 9758 lowscn 569471890 found to restore
RMAN-06025: no backup of log thread 1 seq 9757 lowscn 569440076 found to restore
RMAN-06025: no backup of log thread 1 seq 9756 lowscn 569411439 found to restore
RMAN-06025: no backup of log thread 1 seq 9755 lowscn 569378529 found to restore
RMAN-06025: no backup of log thread 1 seq 9754 lowscn 569358970 found to restore
RMAN-06025: no backup of log thread 1 seq 9753 lowscn 569321882 found to restore
RMAN-06025: no backup of log thread 1 seq 9752 lowscn 569284238 found to restore
RMAN-06025: no backup of log thread 1 seq 9751 lowscn 569261110 found to restore
you can see that something is setting the recovery SCN to 569505448 which higher even then any of the archived logs mentioned above. If I select current_scn from the production database, this gives me 569528258 which is closer to the value which RMAN is expecting to recover to than any of the archived redo logs.
Can you think what might be causing RMAN to try to recover to this value? and why does it appear to be ignoring the SET UNTIL SEQUENCE command?
Cheers,
Phil