Question & Answer
How to recover volumes that have errors or have been marked unavailable.
Tape errors can occur during the normal course of operations on the Tivoli Storage Manager server. This procedure discusses options for recovering from minor or major tape failures.
1) Check the tape status. The tape will be marked 'unavailable' if too many tape errors have occurred or if the Tivoli Storage Manager server lost access to the drive while using the tape. 'QUERY VOLUME <volume> F=D' will show the status and the number of read or write errors.
2) Mark the tape "readwrite" via the "UPDATE VOLUME" command and try the operation again unless you know that the tape is bad. If the tape is known to be bad then skip to step 4. Using a bad tape can damage the tape drive.
3) Try to move the data off of the tape, by first doing an 'UPDATE VOLUME' to make it read-only and then 'MOVE DATA' commands. If this works, you should now have salvaged the data from the tape, and may want to remove this bad tape volume from the library. If the volume shows no utilization and 'QUERY CONTENT <volume name>' shows no files, mark the tape as destroyed (with 'UPDATE VOLUME') and check it out of the library. If it still has data that needs recovery, proceed with the next steps.
At this point if the tape has damaged data that could not be moved, then you must rebuild the data from other copies. Continue with the following Restore Volume procedure to restore the data from your copy pool.
NOTE: You will not be able to use the following procedure if you do not have a copy pool that contains data for the primary pool which contains the damaged volume. If you do not use a copy pool, and the data still exists on the client, it will be backed up again on the next incremental backup after the primary pool volume that contained the bad data is deleted. If you do not use a copy pool, and the data does not still exists on the client, that data will be lost.
If you are unsure if the data on the volume has a copy, you can confirm this with the following command,
- query content <volume name> copied=yes
Any objects that have copies will be listed and you can proceed with the next step (4) to recover those objects. If no objects are listed then at this point the volume should be deleted as the remaining steps do not apply. This can be done with the following command,
- delete volume <volume name> discarddata=yes
4) Make sure that "MOVE DATA" or "AUDIT VOLUME" commands are not running on this volume before proceeding. Also, keep the reclamation threshold at 100% for the copy storage pool during the remainder of the procedure.
5) Update the status of the volume to DESTROYED so that the Tivoli Storage Manager Server will not try to use this tape. You can use the following command:
- UPDATE VOLUME <volume name> ACCESS=DESTROYED.
6) Generate a list of which copy storage pool volumes will be necessary to reconstruct the primary storage pool volume with:
- RESTORE VOLUME <volume name> PREVIEW=YES
7) For each of these volumes, list the current access and mark the volume as unavailable. This will protect the data so it is not reclaimed while in transit (and thus the list changes). Save a copy of the status of each volume so you can restore it when finished.
- QUERY VOLUME <COPYPOOL volume name> F=D
UPDATE VOLUME <COPYPOOL volume name> ACCESS=UNAVAILABLE
8) Bring the needed copy storage pool volumes back onsite and check them into the library using the CHECKIN LIBVOLUME command. Change their access to "readonly" as we do not want them to be reclaimed or have their contents changed.
- UPDATE VOLUME <COPYPOOL volume name> ACCESS=READONLY
9) Restore the TAPEPOOL volume now:
- RESTORE VOLUME <volume name>
10) The volume you are restoring will automatically be updated to an access of "destroyed" if this was not already done manually and a new tape volume will be generated to take its place. When the restore process finishes you may now delete the volume if it still exists in the primary storage pool (using the DELETE VOLUME command with the discarddata=yes option). Any data that remains on the volume after the restore is finished is not recoverable.
11) The COPYPOOL volumes used in the volume restoration can now be changed back to their original access.
- UPDATE VOLUME <COPYPOOL volume name> ACCESS=<original access>
12) Check the COPYPOOL volumes out of the library using the CHECKOUT LIBVOLUME command and return them to offsite. (If using DRM, the volumes will be in MOUNTABLE state and can be moved offsite through normal DRM processing)
A. If you continue to have tape errors, check your hardware for problems.
B. The restore volume command automatically changes the disposition of the original volume to "destroyed". In the above steps, the volume was manually marked as destroyed.
C. You can either delete the damaged volume or let expiration delete it. Either way, the volume should be checked out of the library with CHECKLABEL=NO, or if this fails, with the FORCE=YES parameter.
See the Tivoli Storage Manager Administrator's Reference guide for complete details on these commands.
17 June 2018