Administrator's Guide

Correcting Damaged Files

A data error can be caused by such things as a tape deteriorating or being overwritten or by a drive needing cleaning. If a data error is detected when a client tries to restore, retrieve, or recall a file or during a volume audit, ADSM marks the file as damaged. If the same file is stored in other copy storage pools, the status of those file copies is not changed.

If a client tries to access a file that is marked as damaged and an undamaged copy is available on an onsite copy storage pool volume, ADSM sends the user the undamaged copy.

Files that are marked as damaged cannot be:

• Restored, retrieved, or recalled

• Moved by migration, reclamation, or the MOVE DATA command

• Backed up during a BACKUP STGPOOL operation if the primary file is damaged

• Restored during a RESTORE STGPOOL or RESTORE VOLUME operation if the backup copy in a copy storage pool is damaged

Maintaining the Integrity of Files

To maintain the data integrity of user files, you can:

1. Detect damaged files before the users do.

   The AUDIT VOLUME command marks a file as damaged if a read error is detected for the file. If an undamaged copy is in an onsite copy storage pool, it is used to provide client access to the file.

2. Reset the damaged status of files if the error that caused the change to damaged status was temporary.

   You can use the AUDIT VOLUME command to correct situations when files are marked damaged due to a temporary hardware problem, such as a dirty tape head. ADSM resets the damaged status of files if the volume in which the files are stored is audited and no read errors are detected.

3. Correct files that are marked as damaged.

   If a primary file copy is marked as damaged and a usable copy exists in a copy storage pool, the primary file can be corrected using the RESTORE VOLUME or RESTORE STGPOOL command. For an example, see "Restore Damaged Files".

4. Regularly run commands to identify files that are marked as damaged:

   • The RESTORE STGPOOL command displays the name of each volume in the restored storage pool that contains one or more damaged primary files. Use this command with the preview option to identify primary volumes with damaged files without actually performing the restore operation.

   • The QUERY CONTENT command with the DAMAGED option lets you display damaged files on a specific volume.

For an example of how to use these commands, see "Restore Damaged Files".

Restore Damaged Files

If you use copy storage pools, you can restore damaged client files. You can also check storage pools for damaged files and restore the files. This section explains how to restore damaged files based on the scenario in "Example: Simple Hierarchy with One Copy Storage Pool".

If a client tries to access a file stored in TAPEPOOL and a read error occurs, the file in TAPEPOOL is automatically marked as damaged. Future accesses to the file automatically use the copy in COPYPOOL as long as the copy in TAPEPOOL is marked as damaged.

To restore any damaged files in TAPEPOOL, you can define a schedule that issues the following command periodically:

restore stgpool tapepool

You can check for and replace any files that develop data-integrity problems in TAPEPOOL or in COPYPOOL. For example, every three months, query the volumes in TAPEPOOL and COPYPOOL by entering the following commands:

query volume stgpool=tapepool

query volume stgpool=copypool

Then issue the following command for each volume in TAPEPOOL and COPYPOOL:

audit volume <volname> fix=yes

If a read error occurs on a file in TAPEPOOL, that file is marked damaged and an error message is produced. If a read error occurs on file in COPYPOOL, that file is deleted and a message is produced.

Restore damaged primary files by entering:

restore stgpool tapepool

Finally, create new copies in COPYPOOL by entering:

backup stgpool tapepool copypool