Search this blog

Thursday, January 21, 2010

RMAN error – Use CROSSCHECK command to fix status

Our RMAN backup control mail gave an error: x objects could not be deleted for DISK channel(s) due to mismatched status. Use CROSSCHECK command to fix status.

CROSSCHECK is a check to determine whether files on disk or in the media management catalog correspond to the data in the RMAN repository. Because the media manager can mark tapes as expired or unusable, and because files can be deleted from disk or otherwise become corrupted, the RMAN repository can contain outdated information about backups. Crosschecks update outdated RMAN repository information about backups whose repository records do not match their physical status. For example, if a user removes archived logs from disk with an operating system command, the repository still indicates that the logs are on disk, when in fact they are not. The crosscheck command is used to validate RMAN records in the database control file and the recovery catalog against what is physically on the backup media. The crosscheck command can be used on both disk backups and tape backups. You can cross-check the gambit of backups, from database backups and archive-log backups to image copies, the crosscheck command covers them all.

When you run the crosscheck command, any missing backup files will be marked as EXPIRED, meaning that they are no longer on the media where they are expected to be. The list expired command will show you the backups that are expired. You can review this list and then use the delete command to mark the backup files as deleted in the control file and the recovery catalog. The CROSSCHECK command does not delete any files that it is unable to find, but updates their repository records to EXPIRED. Then, you can run DELETE EXPIRED to remove the repository records for all expired files as well as any existing physical files whose records show the status EXPIRED. Expired backups will not show up on this report until the crosscheck command detects they are missing.

Oracle SID should be correctly set, so we'll log on directly.

F:\oracle\rman>rman target =/

We use show all to get the backup retention policy to determine how long backups and archived logs need to be retained for media recovery. You can define a retention policy in terms of backup redundancy or a recovery window. RMAN retains the datafile backups required to satisfy the current retention policy, and any archived redo logs required for complete recovery of those datafile backups. In our case I got RETENTION POLICY TO REDUNDANCY 1 and ARCHIVELOG BACKUP COPIES FOR DEVICE TYPE DISK TO 1.

RMAN> show all;

RMAN> delete noprompt obsolete;

Will give the same errors as in our control mail, so let’s do what it wants.

    RMAN> crosscheck archivelog all;

RMAN is not removing all of the files because some of them may still be needed for a full recovery!  It's all about how RMAN defines an obsolete file.  Generally, an obsolete file is one that supplements a full backup that will never be used for a recovery and roll-forward. The Oracle docs note the rules for a file becoming obsolete:
DELETE OBSOLETE does not delete backups required to satisfy the specified retention policy, even if some backups have KEEP UNTIL times set which have passed to override the retention policy.
Backups are never obsolete if they are still needed to meet the retention policy, regardless of any KEEP UNTIL time. With a recovery window-based retention policy, even if the specified KEEP UNTIL time has expired, the backup is retained if the backup is needed to satisfy the recovery window.

With a redundancy-based retention policy, even if the specified KEEP UNTIL time has expired, the backup is retained as long as it is required to satisfy the redundancy requirement. 
You can also use the REDUNDANCY or RECOVERY WINDOW clauses with DELETE to delete backups obsolete under a specific retention policy instead of the configured default:
DELETE OBSOLETE REDUNDANCY = 3;
DELETE OBSOLETE RECOVERY WINDOW OF 7 DAYS;

RMAN> delete noprompt obsolete;

I got a similar error as in the controlemail, this time for the controlefilecopy (32 is here the key):

RMAN> crosscheck controlfilecopy 32;

Now we can delete them:

RMAN> delete noprompt obsolete;

And rerun the last delete to verify if anything’s left:

References:
http://download.oracle.com/docs/cd/B19306_01/backup.102/b14192/maint002.htm
http://users.telenet.be/oraguy.be/rman1.htm
http://download.oracle.com/docs/cd/E11882_01/backup.112/e10643/toc.htm

No comments:

Post a Comment