We have two separate locations. Location A is our production facility with several core servers, while Location B has a separate network and another server stack. We want backups from Location A stored in immutable Wasabi cloud storage, while still being able to access those backups from Location B and recover or boot the workloads there during a disaster.
Our original setup used Veeam to back up to a Synology NAS, Hyper Backup to copy the backup data to Wasabi, and then another Synology at Location B to download and open the backup set with Veeam. It worked, but required manual intervention and we later discovered that Hyper Backup does not provide the immutability we need.
Veeam can write directly to Wasabi with immutability enabled, but we have not been able to get the Veeam installation at Location B to use the same bucket for restores. Downloading the files to the Location B NAS and trying to open them there also failed.
What would be a sensible design for maintaining immutable backups while also having a practical recovery or failover option at Location B?
4 Answers
The Veeam server at Location B should be able to connect to the Wasabi bucket as a read-only object storage repository and perform restore operations. You may not want to copy the backup files to the second NAS at all. Also check whether the bucket is configured as a Veeam-compatible object repository and whether the necessary credentials, endpoint, networking, and object-lock settings are available at Location B.
A common approach is to use dedicated backup appliances at both locations. Each appliance keeps immutable local backups and replicates backup data to the other site. That gives you fast local restores, an off-site copy, and a recovery path even if one facility is unavailable. Products such as Rubrik are designed around this model, although the same general architecture can be implemented with other backup platforms.
The current chain of NAS to backup software to cloud to another NAS adds a lot of operational complexity. Another option is to use a cloud backup and disaster-recovery service that backs up production directly, keeps the backups immutable, and can start failed machines in the provider's cloud. A site-to-site VPN can then provide access while the primary location is down. This is easier to operate, but it is a different model from maintaining a warm spare environment at Location B and can be more expensive.
If the goal is to have systems ready to run at Location B, replication is probably a better fit than treating Wasabi as the middleman. Veeam can run replica jobs from Location A to Location B at a suitable schedule and bandwidth limit. Keep the immutable Wasabi backups separately for longer-term recovery and protection against ransomware. Your virtualization platform may also provide its own replication feature, which could be simpler than moving everything through backup storage.
That makes sense. We may need both site replication for a quick failover and independent immutable backups for recovery scenarios that replication alone would not cover.
That distinction is important: spinning workloads up in a provider's cloud is not the same as having a prepared secondary site with local servers. The right choice depends on the required recovery time, application dependencies, and whether Location B must remain operational during a prolonged outage.