One of the powerful features of Veeam Backup & Replication is its ability to perform incremental backup jobs to local disk, and then schedule jobs to copy those increments to another location for additional protection.
Backup Copy Jobs: Forever Forward Incremental
Veeam’s backup copy jobs operate using a forever forward incremental method. After the first full backup is created, only incremental changes are copied moving forward. Once the incremental chain reaches the set limit of restore points, the oldest restore point is automatically merged into the full backup file after a new restore point is added.
By default, Veeam keeps 7 incremental restore points, but this number can be adjusted. Managecast, for instance, increases this to 14 daily incrementals by default, allowing for two weeks of daily recovery points.
Long-Term Retention: GFS (Grandfather-Father-Son)
For longer retention, Veeam recommends using GFS retention (Grandfather-Father-Son) within backup copy jobs. GFS allows you to retain a set number of weekly, monthly, quarterly, and yearly full backups.
With GFS, a backup copy job creates a new full backup file and archives it based on your retention policy. These GFS restore points are independent full backups, meaning they don’t rely on the incremental chain. This is an advantage, as they won’t be affected if the incremental chain is broken.
Storage Considerations for Long-Term Retention
While GFS provides robust retention capabilities, it can also quickly consume a lot of storage space. For example, if you retain 7 incremental restore points, 4 weekly backups, and 12 monthly backups, you’ll need storage for:
- 7 incrementals
- The current full backup file
- 15 additional full backup files
A helpful resource to estimate storage needs is the restore point simulator.
Reducing Storage Usage with Deduplication
One way to cut down on storage requirements is to use deduplication on the target backup repository. Because GFS restore points are copies of the same backup files, they deduplicate efficiently.
However, deduplicating the current full backup can significantly slow down the merging process for the oldest incremental restore point. To avoid this, Managecast only deduplicates files that are older than 7 days. This ensures that GFS restore points are only deduplicated a week after being copied, leaving daily incrementals and the current full backup untouched by deduplication until then.
With this approach, the repository stores daily incrementals, the current full backup file, and deduplicated GFS full restore points. This setup typically requires just over 2x the full backup size plus the size of the incrementals.
Important Consideration: Encryption and Deduplication
It’s important to note that deduplication will not work on encrypted files. Encryption changes the file contents, making them unique and preventing deduplication from recognizing similarities across backup files.
This presents a choice:
- Encrypt the backup files and have a higher-cost long-term retention policy, or
- Forego encryption in favor of a deduplicated long-term retention strategy.
Summary
Veeam is an excellent product, but long-term retention requirements can quickly increase storage needs. At Managecast, we are continuously exploring new technologies to address these challenges. We’re currently reviewing Veeam v9.5 alongside Windows 2016 and the ReFS file system to evaluate if new deduplication efficiencies (with combined encryption) can help solve these storage issues. Stay tuned for more updates!
Check out our post on Veeam 9.5 ReFS Integration for more details on long-term retention improvements!
