Infrastructure
Testing a recovery plan across multiple sites.
The client needed to recover data across several sites if a backup location failed or an attacker reached its systems. I designed the backup architecture around the copies its operators would actually use in a recovery.
Building for a loss of site
Copies held in one place can share the same point of failure. The design combined fast local recovery with copies held elsewhere and a retention period during which backup data could not be altered.
We used the 3-2-1-1 approach. It calls for three copies across two types of media, with one copy off site and one offline or immutable.
Where each copy lives
Veeam writes to hardened Linux repositories using the XFS file system. These repositories hold an immutable copy during the configured retention period. Dell PowerProtect Data Domain uses DD Boost to reduce duplicate data written to storage.
Cross-site replication gives the team another recovery location. An off-site copy in Wasabi object storage adds a further option if the local infrastructure is unavailable.
Testing the route back
We ran restore tests across sites and documented the recovery steps. That gave the operations team a tested sequence to follow and a way to check whether changes to the environment affected recovery.