Best Practices for Large Tivoli Storage Manager Client Restores in the Event of Device Failures

The following solution assumes that the system that suffered the data loss was first registered with the Division of Information Technology's (IT) Data Backup Service.  It also assumes backups were taken of the system in question.  If in doubt, contact the Call Center.

In cases where there has been a catastrophic failure or a large amount of data-loss and you would like to restore data from the Backup Service, open a support ticket with the Division of IT to alert backup administrators of a pending large restore (e.g. where an entire file system is lost).  Division of IT administrators will work with you to ensure you have access to server resources for your restore.

Here are some technical hints on a large restore:

Use the command line because it is easier to see what the restore is doing, and because you can cancel out the restore a little faster.  If you see the restore is not going the way you want within the first minutes, you can cancel, recall the command, edit, and try again.  It may take a few tries to get the restore going exactly the way you want it.

To get command-line in Windows environments:

1. Open cmd windows as an Administrator.

2. cd to C:\Program Files\tivoli\tsm\baclient\ 

3. Enter dsmc (enters you into the Tivoli Storage Manager (TSM) command-line. My sample commands execute dsmc in batch mode, meaning valid dsmc commands are added as parameters to the command-line).

To get command-line in Macintosh:

1. Open Terminal.

2. Issue sudo su, and enter your system password.

3. Run dsmc (enters you into the TSM command-line.  My sample commands execute dsmc in batch mode, meaning valid dsmc commands are added as parameters to the command-line).

For large restores reference this documentation:

Also consider using restartable restore sessions:

NOTE: If you issue commands that generate lots of restartable restores, you may run into errors that say a 'restartable restore is pending'.  In which case, you can issue query restore and cancel restore commands to remove any stale restartable restores.

Before beginning, you can multithread the restore sessions by increasing the resourceutilization parameter.  Set it to 3:

And this table lists some restore command options:

In most situations where you just want to restore everything from a drive or mountpoint, run with something similar to this:

dsmc restore -resourceutilization=3 'D:\oldpath\*' 'D:\restorepath\' -subdir=yes

dsmc restore -resourceutilization=3 '/oldpath/*' '/restorepath/' -subdir=yes

NOTE: This command assumes you just want to restore everything from the latest backup under D:\oldpath or /oldpath directory.  If you need to do a point-in-time, look in the documentation for pitdate and pittime options.  



From the command-line, you can pipe the output to a file and open it with text edit to see exactly what was restored.  From the GUI, this is not possible.