Block-Level Restores

You can perform a file-level or a volume-level restore from a block-level backup.

Remember: You cannot apply filters for block-level restores.

High-Level Process for Block-Level Restores

  1. If you are performing a volume-level browse and restore, the entire volume is restored.

  2. If you are performing a file-level browse and restore, the following operations occur:

    • Depending on whether the metadata collection was enabled during backup, one of the following browse methods is selected if you perform a browse operation:

      • If metadata collection was enabled, a browse from the index is initiated.

        Note: We recommend that you use the metadata collection option with caution because it might affect block-level backup performance.

      • If metadata collection was disabled, a live browse is initiated.

    • The selected files and folders are restored.

Live Browse

Applies to: Microsoft Windows 2008 R2 or later.

Live browse initiates a pseudo-mount which constructs a device to store the blocks. The size of the staging folder must be at least 4 GB. The recommended size of the staging folder is 10% of the restore data, and it can be as large as the restore data. If the size of the staging folder is smaller than the restore data, then the extents are pruned based on the LRU (Least Recently Used) algorithm, for the new extent restores. This might result in decreased performance during the restore operation.

For more information on configuring live browse, see Configuring Live Browse for Block-level Backups on Windows Computers.

Remote Browse

In the Browse and Restore dialog box, Advanced Options tab, the client computer is selected in the Use proxy box, by default. However, if a different MediaAgent that is not compatible with the operating system is chosen, then the browse operation might fail.

For example, if the Windows client computer is backed up to a Linux MediaAgent and the Windows client computer is chosen as proxy MediaAgent, then you can browse the files.

Multi-Stream and Multi-Node Restores

By default, block-level restore operation uses a single stream. For faster restore operations, you can configure the restore operation to use multiple streams, and you can define alternate data paths. For more information, see GridStor® (Alternate Data Paths) - Overview.

By default, block-level restore operation uses a single node when you perform restores to CIFS shares (UNC path). For faster restore operations, you can configure the restore operation to use multiple nodes, and you can define alternate data paths. For more information, see GridStor® (Alternate Data Paths) - Overview.

If you want to perform restores to distributed file systems or to network file systems (such as CIFS) that are mounted across multiple computers, use multiple nodes.

Notes:

  • You cannot use multiple streams and multiple nodes for restore operations on clients with only File System Core package installed (Laptop Backup clients).

  • You cannot use multiple streams and multiple nodes for restoring systems state backups.

Pruning Blocks to Reclaim Space

The restored blocks are saved in the device files under the job results folder. To reclaim space in the job results directory, the restored blocks are pruned using the LRU (Least Recently Used) algorithm.

Best Practice

  • Stage the extents on a high-performing disk storage media (SSD).

  • Perform file system defragmentation on the source volume.

Loading...