Veeam Offload to Object Storage: How It Works

Veeam Backup & Replication can offload backup data to object storage (like Amazon S3, Azure Blob, or S3-compatible on-premises appliances) using its Scale-out Backup Repository (SOBR) tiering feature. You configure a capacity tier and set a policy that moves older backup files (or copies of them) from fast local storage to the object store, which cuts storage costs while keeping backups restorable. This article explains what offloading actually does, how to set it up, what the common pitfalls are, and how it differs from simple archive copies.

What Object Storage Offloading Means in Veeam

Veeam’s SOBR is a logical repository that spans multiple storage extents. You can add an object storage bucket as a capacity tier. Once configured, Veeam can move backup files from the performance tier (local disk or NAS) to the capacity tier based on age or when the performance tier reaches a certain occupancy level. The data is stored in a proprietary format, not as plain files, and Veeam maintains metadata in the local repository so you can browse and restore without downloading the entire backup.

There are two main modes: tiering (move) and copy. In move mode, the backup file is relocated to object storage and removed from the performance tier. In copy mode, a second copy is placed in object storage while the original remains. You can also enable archive tier for Amazon S3 Glacier or Azure Archive, which moves data to even cheaper cold storage after a second retention period.

How the Offload Process Works

When you create a SOBR and add a capacity tier, Veeam runs a background job that scans the performance tier for backup files that meet your policy. The policy options are:

• Age-based: Move backups older than N days (e.g., 30 days).
• Space-based: Move the oldest backups when the performance tier usage exceeds a percentage (e.g., 80%).

Once a file is selected, Veeam performs the following steps:

1. It reads the backup file blocks from the performance tier.
2. It encrypts the data if you enabled encryption at the SOBR level.
3. It uploads the blocks to the object storage bucket using the S3 API (or Azure Blob API).
4. It updates its internal database to record where each block lives.
5. It deletes the local copy (in move mode) after a successful upload.

Restores read directly from object storage, so you don’t need to stage the data back to local disk first. However, for large restores, performance may be slower than local disk, especially with cold or archive tiers.

Choosing the Right Object Storage Provider

Veeam supports any S3-compatible object storage, which includes major public clouds and many on-premises appliances. When choosing, consider:

• API compatibility: Veeam requires certain S3 features (like multipart upload and object locking for immutability). Most modern providers support these, but some older or minimalist appliances may not.
• Immutability: For ransomware protection, enable object lock on the bucket. Veeam can then make backups immutable for a set period, preventing deletion even by an admin.
• Performance tier vs. archive tier: If you use move mode, restores from archive can take hours because the provider must first retrieve the data. Keep recent backups in the performance tier or use copy mode for faster access.
• Egress costs: Public cloud providers charge for data transfer out. If you restore often, these costs can offset the savings from cheaper storage.

Common Problems and Misconceptions

One common misconception is that offloading to object storage is the same as making a backup copy. It is not a backup copy—it is a tier of the same repository. If you accidentally delete a backup file from the performance tier before it is offloaded, the data is gone. Always maintain a separate backup copy (e.g., Veeam Backup Copy job) for disaster recovery.

Another issue is that Veeam does not support direct restores from archive tier in all versions. If you move data to Glacier or Azure Archive, you must first retrieve it to the standard tier, which can take hours. Plan your retention so that only very old backups end up in archive.

Performance problems can also arise if the object storage is far from your Veeam server or if the network is slow. Veeam uses multiple concurrent connections, but high latency can still slow down backups and restores. For on-premises S3-compatible storage, place it on the same LAN as your Veeam server.

Costs and When It Makes Sense

Object storage is typically cheaper per GB than high-performance local storage, especially for cold tiers. For example, a 100 TB backup repository on local disk might cost several thousand dollars in hardware, while the same data in S3 Standard costs less per month, and even less in Glacier. However, you must factor in egress fees, API request costs, and the cost of retrieving data for restores. Offloading makes sense when:

• You need long-term retention (months or years) and rarely access old backups.
• You want to reduce the size of your local storage footprint.
• You need off-site or geo-redundant storage for disaster recovery.
• You want immutable backups for ransomware protection.

It does not make sense if you need instant restores of every backup, or if your object storage is slower than your local disk and you have no budget for egress.

Alternatives to SOBR Offloading

If you don’t want to use SOBR, Veeam offers other ways to get data to object storage:

• Backup Copy Job to object storage: You can create a backup copy job that writes directly to an object storage repository. This creates a separate, independent copy, not a tier.
• Archive Repository: Veeam can use object storage as an archive repository for long-term retention, but this is a manual or job-based process, not automatic tiering.
• Veeam Cloud Connect: If you use a service provider, they may offer object storage as part of their service, but the configuration is on their side.

Frequently Asked Questions

Does offloading to object storage free up space on my local backup repository?

Yes, in move mode, Veeam deletes the local copy after the data is safely uploaded to object storage. This frees up space on the performance tier for new backups. In copy mode, the local copy remains, so no space is freed.

Can I restore a VM directly from object storage?

Yes, Veeam can restore directly from object storage without staging the data back to local disk. However, the restore speed depends on the object storage performance and your network connection. For large VMs, it may be slower than restoring from local disk.

What is the difference between capacity tier and archive tier?

Capacity tier is the primary object storage tier (e.g., S3 Standard). Archive tier is a second, colder tier (e.g., S3 Glacier) that you can enable after data has been in capacity tier for a set period. Archive tier is cheaper but has longer retrieval times.

Does Veeam support immutability with object storage?

Yes, if your object storage provider supports object lock (S3 Object Lock or Azure Blob immutability), Veeam can make backups immutable. This prevents modification or deletion until the lock expires, which is a strong defense against ransomware.

Can I use any S3-compatible storage with Veeam?

Veeam supports any S3-compatible storage, but you should check the provider’s compatibility with Veeam’s required features. Some providers may not support object lock or multipart upload correctly, which can cause issues.

What happens if the object storage is temporarily unavailable?

If the capacity tier is unreachable during an offload job, Veeam will retry the job according to its retry policy. If it fails permanently, the data remains in the performance tier, and the SOBR will continue to operate, but you may run out of space if the offload never completes.

Related

• Solana Treasury Shrinkage: What It Means
• Mac Won’t Go to Sleep? Fixes That Actually Work
• Violet Abilities & Beginner Playstyle in Deadlock