Using Mirrored Restore to Prepare a New Server

Administrators may require a restore method to prepare a new file server for production service that is synchronized with the original server's data prior to switchover to the new server. 

The time required to restore a huge data set may exceed the time available for the restore job to complete. In this case, two restores can be performed, the first to restore the bulk of the data well in advance of the date to put the new server in service , and the second just prior to putting the new server in service that uses a mirrored restore of the latest restore point to update the restore path with new and changed files and remove extras that do not exist in the latest restore point. The mirrored restore should finish in time because most of the data in the restore path will not have changed when compared to the prior restore job. A mirrored restore also provides an exact copy of the latest restore point.

Prerequisites: Make sure you have installed build 9.0.2703.1 or later of the software to support mirrored restore. The Alerts button provides a download link for the latest build of the software when one is available. Contact support if you need assistance obtaining the latest build.

Option one: If you can move the backup storage to the new server for the restore process, or in the case of a network drive if you can access it from there, follow these steps:

  • Install the DATASTOR software on the new server. Add the permanent key but do not attempt to activate it. Click 'remind me later'. At this point you will have 30 days of access to the software without activating the license. Refer to Gaining 30 Day Access to DATASTOR Shield Professional Software
  • Move the RDX dock or other storage to the new server.
  • Use the Storage button to add storage. Point to the storage device to attach to existing storage.
  • Click the Explore button. In the View Options table, uncheck the box to only show recovery points for this server. Then you can see the restore points generated from the original server.
  • Run an initial restore job of the data set you want to restore to the new server, using the most recent successful backup restore point.
  1. If you wanted to restore all folders on the old server's D: drive to the new server's E: drive, you would highlight the D: drive, right click and choose restore from the pop up menu.
  2. Select to restore to an alternate location. Browse to the E: drive and select it as the restore path.
  3. On the first restore job, choose to restore missing files, which will restore all files because the E: drive will be empty, presumably.
  4. Check the box to restore with high performance.
  5. Click the Restore button to start the job.

Once the restore on the new server completes, move the storage back to the original server. Run the protection plan on the original server just prior to the second restore job to generate another restore point. Then, move the RDX dock to the new server again.

Run a mirrored restore using the same source path and the same restore path you used on the first restore job, using the latest restore point.

  1. Explore the latest restore point. Right click the D: drive and choose 'restore' from the contextual menu. Select to restore to an alternate location. Browse to the E: drive and select it as the restore path.
  2. On this restore job, choose from the drop down list to restore "missing files, files that have changed, and remove extras", which will mirror the archive restore point exactly, beginning within the top level folders that restore, e.g. E:\dir1, E:\dir2 and so on and their subfolders and files. Any existing files and folders at the root of the E: drive are not removed.
  3. Check the box to restore with high performance.
  4. Click the Restore button to start the job.

If the amount of changed data between the two restore points was only 50 GB, then only 50 GB will be written to the restore path, even though the entire data set in the restore point may be multiple Terabytes of data.

Option two: If you cannot move the software and storage to the new server for the first restore job, because you need to continue running backups, for example, then you could restore from the original server across the network, though it would be slower. 

Map a drive as an administrator to the restore path so it appears in the list of available restore locations (under the "This PC" folder). Follow the steps in our kb article Restoring to an alternate network location to prepare the network location. Then follow the two restore jobs outlined above. 

Note: Make sure the restore point was generated successfully by the protection plan prior to using it for the mirrored restore. Review the results and the plan log. The mirrored restore removes extra files from the restore path that are not present in the archive restore point.

Note: take care to restore from the same folder level in the restore point to the same folder level in the restore path during both restore jobs, to ensure mirroring works as expected. That is, in the example presented restore missing files from D:\ to E:\ on the first restore job, then restore missing files, files that have changed, and remove extra files from D:\ to E:\ on the second restore job.

Note: a restore job replicates the entire directory tree of the folder selected for restore, not just the folder selected. So, for example, if you explored the restore point and browsed to D:\Users\Gregory and highlighted the folder Gregory, right clicked and selected restore, then the restore job would create the entire Users\Gregory directory tree at the restore path selected. If the restore path selected was E:\Users, the end result would be E:\Users\Users\Gregory. It is likely that you would want to select the restore path E:\ so the result was E:\Users\Gregory.