ORA-19566: exceeded limit of 999 corrupt blocks for file

Question

Hi All,
I am new to Oracle RMAN & RAC Administration. Looking for your support to solve the below issue.
We have 2 disk groups - +ETDATA & +ETFLASH in our 3 node RAC environments in which RMAN is configured in node-2 to take backup. We do not have RMAN catalog and the RMAN is fetching information from control file.
Recently, the backup failed with the error ORA-19566: exceeded limit of 999 corrupt blocks for file +ETFLASH/datafile/users.6187.802328091.
We found that the datafiles are present in both disk groups and from the control file info, we got to know that the datafiles in +ETDATA are currently in use and +ETFLASH is having old datafiles.
RMAN> show all;
RMAN configuration parameters for database with db_unique_name LABWRKT are:
CONFIGURE RETENTION POLICY TO RECOVERY WINDOW OF 3 DAYS;
CONFIGURE BACKUP OPTIMIZATION ON;
CONFIGURE DEFAULT DEVICE TYPE TO DISK; # default
CONFIGURE CONTROLFILE AUTOBACKUP ON;
CONFIGURE CONTROLFILE AUTOBACKUP FORMAT FOR DEVICE TYPE DISK TO '%F'; # default
CONFIGURE DEVICE TYPE DISK PARALLELISM 4 BACKUP TYPE TO BACKUPSET;
CONFIGURE DATAFILE BACKUP COPIES FOR DEVICE TYPE DISK TO 1; # default
CONFIGURE ARCHIVELOG BACKUP COPIES FOR DEVICE TYPE DISK TO 1; # default
CONFIGURE MAXSETSIZE TO UNLIMITED; # default
CONFIGURE ENCRYPTION FOR DATABASE OFF; # default
CONFIGURE ENCRYPTION ALGORITHM 'AES128'; # default
CONFIGURE COMPRESSION ALGORITHM 'BASIC' AS OF RELEASE 'DEFAULT' OPTIMIZE FOR LOAD TRUE ; # default
CONFIGURE ARCHIVELOG DELETION POLICY TO NONE; # default
CONFIGURE SNAPSHOT CONTROLFILE NAME TO '+ETFLASH/CONTROLFILE/snapcf_LABWRKT.f';
CONFIGURE SNAPSHOT CONTROLFILE NAME TO '+ETFLASH/controlfile/snapcf_labwrkt.f';
Above configuration shows that SNAPSHOT CONTROLFILE is pointing to +ETFLASH. So I changed the configuration with the SNAPSHOT CONTROLFILE points to '+ETDATA/controlfile/snapcf_labwrkt.f'. At the end of backup, SNAPSHOT file was created in +ETDATA and I was expecting it to be the copy of control file being used which has dbf located in +ETDATA. But still the backup was pointing to old datafiles in +ETFLASH. Since we dont have RMAN catalog, resync also is not possible.
When I ran it manually, it was successfull without any error and was pointing to the exisiting datafiles.
RMAN> backup database plus archivelog all;
I hope the issue will get resolved if the RMAN points only to the datafiles present in +ETDATA. If I am correct, please let me know how can i make it possible? Also please explain me why the newly created snapshot file does not reflect the existing control file info?

Answer 1

Hi,
I am getting an error from the DBA Planning Calendar every time the job ...
So when was your last successfull backup of this datafile. Check if still available.
If this is some time ago, and may be you are currently without any backup, try to backup without rman at once,
to have at least something to work with in case you get additional errors right now.
Then you need to find out what object is affected. You are on the right way already. You need the statement,
that goes to dba_extents to check what object the block belongs to.
Has the DB been recovered recently, so the block might possibly belong to an index created with nologging ?
(this could be the case on BW systems).
If the last good backup of that file is still available and the redologs belonging to this backup up to current time are as well, you could try to recover that file. But I'd do this only after a good backup without rman and by not destroying the original file.
If the last good backup was an rman backup, you can do a verify restore of that datafile in advance, to check if the corruption is really not inside the file to be restored.
Check out the -w (verify) option of brrestore first, to understand how it works.
(I am not sure it this is already available in version 7.00, may be you need to switch to 7.10 or 7.20)
brrestore -c -m /oracle/SHD/sapdata4/sr3_16/sr3.data16  -b xxxxxxxx.ffr -w only_rmv
You should do a dbv check of that file as well, to check if it gets more information. I.E if more blocks are
affected. rman stops right after the first corruption, but usually you have a couple of those in line, esp. if these are
zeroed ones. (This one would also work with version 7.00 brtools)
brbackup -c -u / -t online -m /oracle/SHD/sapdata4/sr3_16/sr3.data16 -w only_dbv
Good luck.
Volker