How Block-Level Backups on UNIX Work

Specifying the Volume to Back Up

To perform block-level backup operations on a snapshot-capable volume, configure the volume as subclient content. The volume that you back up can be a from a local disk or from a hardware snapshot engine that is supported by Commvault.

The Phases of Block-Level Backup Operations

Block-level backup operations have the following phases:

• Snapshot copy phase
• Backup copy phase

Both of these phases have several sub-phases. For native (CVBF/LVMs) backup operations, both the phases run as a single job. For hardware snapshots, the snapshot copy phase and the backup copy phase run separately as two jobs.

Snapshot Copy Sub-Phases

1. Backup sub-phase: A snapshot is created. For native disks, CVBF snapshots are taken. The snapshots are created in the /dev/CVBF directory. For native LVMs, Commvault uses the LVM snapshot capability for snapshot creation. For hardware backups, Commvault uses the vendor-specific snapshots. The CVBF driver monitors the device and creates a bitmap by tracking the changed blocks while writes are being performed on the volume being monitored. A block device config file is also created which holds the volume details and properties.
2. Cleanup sub-phase: The temporary files that were created under the Job Results directory are removed.

Backup Copy Sub-Phases

1. Scan sub-phase: The volume bitmaps and the block device config file are downloaded from the index. Extent collect files that have data-only blocks are generated.
2. Backup sub-phase: The snapshots are mounted, and the extents are moved from the snapshot to the media.
3. Archive index sub-phase: The index for the backup job is backed up to the media.

For native backup operations, a spool copy is automatically configured, and the snapshots are deleted after the backup copy job completes. For hardware backups, the snapshots are deleted according to the storage policy settings.

File Indexing After Block-Level Backup Job Completion

If you enable the Index files after backup option, files are cataloged offline as a separate job after the block-level backup completes.