In this eLearning module, you will learn about repairing virtual volumes for the IBM® TS7700 family. You will learn how to view damaged virtual volumes, what the potential causes are, how to access and work with virtual volumes, and how to repair them.
Damaged volumes typically occur due to user intervention. In one scenario, a user could have two independent clusters run to the same volumes with ownership takeover enabled. In another scenario, a user does a cluster removal and the consistency check prior to the removal showed that the only valid copies present were on the box being removed.
When a virtual volume is damaged, due to one of the scenarios indicated, the volume will be made inaccessible to the host. You must resolve the conflict by either choosing which volume copy should be used as the working copy or allow the volume to be moved back into a scratch pool for reuse in order to permit future host access.
In the event a volume enters the damaged state because two conflicting consistent versions are found, or if only older versions are available, part of the repair process can include the ability to choose a version and mark it as the most consistent copy.
The virtual volume can become out of sync when an ownership takeover procedure is manually invoked. This can happen when ownership is transferred from a cluster that is still running, and updates occur from the new and original owner during the outage.
Any host initiated access to a damaged volume will be denied and an appropriate error code will be surfaced for the failed host operation. In addition, each time a volume enters the damaged state, a host console message will be surfaced stating that one or more volumes have entered the damaged state.
On the Repair Virtual Volumes panel, a user can see whether a specific VOLSER appears on the Damaged Virtual Volumes table.
You cannot access a virtual volume that is damaged, so you should repair all of the damaged virtual volumes. You can repair up to 10 at a time, and the 10 that you select must all be repaired using the same repair policy.
A cluster must be chosen within the repair policy indicating which cluster copies of the virtual volume you want to begin using when it repairs the damaged virtual volumes selected. You can view details about a specific damaged virtual volume by clicking the name of the volume.
Clicking on the volume name opens the Damaged Virtual Volume Details panel. The Damaged Virtual Volume Details panel provides information about the different cluster instances of the virtual volume, including the last-modified date. Often the cluster with the most recently modified copy is the best selection for the chosen cluster for your repair policy, but the information on this panel can help you decide.
After you select the cluster instance for the volumes you selected to repair, you have to decide between two options: Move to insert category keeping all data or Move to insert category deleting all data. This choice determines whether the data on the chosen cluster virtual volumes is kept or removed. If you no longer need the data on the virtual volumes and simply want to clear the damaged state of the volumes allowing them to be used for scratch, you can choose the delete option.
Another situation where you might need to delete the data is when you are unsure of which cluster instance to use. Selecting Move to insert category deleting all data returns empty virtual volumes with the same name and details to the insert category allowing the tape management system to move them back to scratch.
If you still need the data within the selected volumes, select Move to insert category keeping all data. Whichever data exists on the chosen cluster will be viewed as valid. If no data exists on this cluster for any of the volumes, these volumes will be viewed as valid empty volumes. As a reminder, when choosing more than one volume to repair, the same chosen cluster and repair policy option will be used for all selected volumes.
After selecting the repair policy, the instructions for the data, and the virtual volumes to be repaired by that policy, you can start the repair procedure by selecting Repair from the Select Action list and clicking Go. A confirmation message appears at the top of the panel. Click OK to proceed.
Click View Tasks to be taken to the Tasks panel. Here you can see the progress of the repair request.
Entries listed at the top of the table are either currently in progress, or are scheduled to be performed in the future. You can view the progression of a task by looking at its progress bar. The Status field also lists the total time a task is taking.
The table also lists tasks that have either been completed or failed before being completed. Use this table to view information about and a history of repair processes.
This has been a brief overview of repairing virtual volumes for the TS7700 family. For additional information, see the other topics in IBM Knowledge Center.