This content has been marked as final. Show 2 replies
You could try to:
1. physically move the backups sets to new location with Windows
2. flag as expired all backup sets that RMAN cannot find:
3. uncatalog backup sets from step 2 with:
4. catalog backup sets new location:
DELETE EXPIRED BACKUP;
Note that there is always some risk with DELETE EXPIRED to delete backups that are only temporarily unavailable: you must double check backup sets that are flagged as expired.
CATALOG START WITH <new location>;
Thanks for the reply. Yes, that is pretty much what I did. I first backed up my backups so I would have them just in case. Then ran the following:
I then ran maintenance against the rman catalog by issuing the following. (not completely necessary, but a little overkill here was not going to hurt anything)
Delete Expired Copy;
Delete Expired Backup;
These two steps deleted all the backups I had at the time and any reference to them in the catalog. Then I ran the backup again, and all files were written to the new SIF share. There are some curious differences, however. First, the datafile copies were placed in the root of the SIF share rather than root/dbname/datafiles as previously had occurred. The autobackups and backupsets are still maintaining the same directory structure, i.e., root/dbname/autoback and root/dbname/backupset.
Second, on my separate backup job that I run to capture archivelogs, each time that job runs it is creating another autobackup. Not that this is a bad thing, just different than RMAN used to behave, and I am just not sure why it is behaving differently.