Backup Chain Formats

When you add repositories to the backup infrastructure, you can select how to store backup files. Veeam Backup & Replication provides three backup chain formats:

• Per-machine backup with separate metadata files
• Per-machine backup with single metadata file
• Single-file backup

The backup chain format depends on the Use per-machine backup files option in the repository settings and also the version of Veeam Backup & Replication when the repository was added. You can find the details further in this section.

Per-Machine Backup with Separate Metadata Files

This way of storing backups is the default option for newly added repositories starting from Veeam Backup & Replication version 12. The per-machine backups with separate metadata files are created in backup repositories for which the Use per-machine backup files option is selected.

When per-machine backup with separate metadata files format is used, a backup job writes data in a separate stream for each workload. Veeam Backup & Replication saves data of each workload into a separate backup file and also creates a separate metadata file (.VBM) for each workload. Veeam Backup & Replication perceives each backup file created during one job session as one restore point. As a result, each workload has its own independent backup chain.

Such way of storing backups is recommended and makes the environment management more flexible. Veeam Backup & Replication can perform operations with individual workloads, for example, move a workload and its backups from one job to another, or launch active full backups for individual workloads.

Per-Machine Backup with Single Metadata File

This way of storing backups is available for Veeam Backup & Replication prior to version 12. Starting from Veeam Backup & Replication version 12, you cannot enable per-machine backup with single metadata file. However, Veeam Backup & Replication still creates backups in the per-machine backup with single metadata file format for the backup chains existing on repositories added to Veeam Backup & Replication version prior to 12 and for which the Use per-machine backup files option was enabled.

When per-machine backup with single metadata file format is used, a backup job writes data in a separate stream for each workload. Veeam Backup & Replication saves data of each workload into a separate backup file. However, Veeam Backup & Replication creates a single metadata file (.VBM) for all workloads and perceives all backup files created during one job session as one restore point.

Such way of storing backups is more efficient than single-file backup format. Compared to per-machine backup with separate metadata files format, Veeam Backup & Replication cannot perform some operations with individual workloads. These operations include moving backups, launching active full for individual workloads and so on.

Single-File Backup

This is the default way of storing backup files for Veeam Backup & Replication prior to version 12. The single-file backups are created in backup repositories for which the Use per-machine backup files option is not selected.

When single-file backup format is used, a backup job writes workload data to a repository in one write stream. Veeam Backup & Replication saves data of all workloads into the same backup file and creates one metadata file (.VBM). Such behavior is not optimal if the target storage device is able to write data in multiple streams simultaneously. In this situation, the backup repository may become the bottleneck for the data transfer, even though its resources will not be fully utilized.

Limitations and Considerations for Per-Machine Backup Files

When planning to use per-machine backup files, consider the following limitations:

• We recommend that you use the Use per-machine backup files option, especially for deduplicating storage appliances used as a backup repositories. Veeam Backup & Replication will write machine data to the backup repository in several streams, which will improve the backup job performance. However, deduplication may be less effective compared to deduplication for single-file backups. Deduplication works within one backup file. For single-file backups, one backup file stores data for all machines, that is why data is deduplicated for all machines. Per-machine backups store data for one machine, that is why data is deduplicated for one machine only.
• [For Veeam Agent backup jobs] Veeam Backup & Replication ignores the Use per-machine backup files option. The way how Veeam Backup & Replication creates backup files depends on the objects included in the backup job:

• If several Veeam Agent machines are included in the backup job, Veeam Backup & Replication creates a separate backup file for each machine.
• If failover clusters are included in the backup job, Veeam Backup & Replication creates a separate backup file for each cluster.

Changing Backup Chain Formats

Changing the Use per-machine backup files option for repositories does not affect the way how backups are created. To change the backup chain format, follow the instructions provided in Upgrading Backup Chain Formats.

Related Topics

• Retention Policy for Per-Machine Backup Files

• Adding Backup Repositories
• Upgrading Backup Chain Formats
• Adding Scale-Out Repositories