RMAN control-file-based backups
I am still learning about RMAN so I used the 'consultants' backup script he left behind on one box since it looked good enough ( I arrived after his exit ). I actually put it on other machines and got the backups going there too.
Before moving on to the repository I wanted to know that these control-file backups can be counted on.
The restore to another Sun box was complicated by totally different file structures, no two boxes look the same here.
But! Oracle tolerated the soft links quite well.
I restored spfile to pfile
altered it considerably
created spfile from pfile
restored control file
restored all datafiles
Could not open the database system.dbf is not sufficiently old.
3 Questions for you RMAN experts
1) Can you only rename datafiles if you are NOT using a repository? Softlinks worked sufficiently as a workaround.
2) Redo logs? Since online redo logs are not backed up by RMAN but already exist in the control file should I drop them and add them back to recreate them on this other host before attempting to open the database?
3) The open database failed, says system.dbf is not sufficiently old.
I think the problem is that his script backsup vital components in the wrong order - namely *.dbf,spfile,controfile,archivelogs
So when you 'list backup' on the restore machine it does not know about the backed up archive logs ( having been backuped up after the control file ) and you cannot recover all the way.
Seems prudent to backup the control file last since that is where the backups are being recorded....but the error message actually sounds opposite to me, I would think my system.dbf is 'too' old.
So perhaps I have another problem, database is small, the backup set is done in less than 5 mins so other than the sequence of events I would think its a compatiable set for a full recovery.
Apparently my failing remedies accomplished something last night because Oracle Support wanted to see the full stack and it had srolled off so I again did submitted..
restore database ;
recover database noredo ;
alter database open resetlogs;
on the disaster recovery host and - IT WORKED.
Weird that just begs more questions......
Click Here to Expand Forum to Full Width