Table of Contents
Overview
One thing many organizations do not consider when creating their BCDR plan is how to restore workloads back to their production environment once they are running in Cyber Fortress' infrastructure. This article will cover the most commonly used methods.
Method 1 - Replication Failback
With replication, Veeam Backup & Replication offers an easy solution: Replica Failback, a multi-stage process. In stage 1, changes are synchronized back to production. During this stage, the replica VMs are still accessible to end users. This synchronization can occur with the original production VM, a production VM restored from backups, or a production VM newly created by Veeam during fallback initialization.
During stage 2, processes are shifted from the replica VM back to production, and the replica VM is powered off. After this, another transmission including all data changed during stage 1 is performed. This second stage can be performed automatically, at a scheduled time, or manually to allow for failback to perform during preferred windows and ensure application consistency during the process.
Stage 3 is the final stage and should be performed as close to the completion of stage 2 as possible. During stage 3, failback is either committed and operations remain in production, or failback is reverted and workloads transition back to the replica VM.
For more information on how failback works, you can read Veeam's official user guide.
Method 2 - Reverse Seed Load
Reverse seed load is most commonly used for workloads protected only by Cloud Connect backup, but can also be used in situations where replica VMs are large and customer bandwidth is not expected to allow for a small transmission time. In a reverse seed load, a backup chain of the restored workloads is sent to the customer site, either using the Cloud Connect protocol or a seed device.
Once on-premises, a final incremental is then transmitted to incorporate all changes made during transit. Then, a blackout window is opened where the restored workloads are powered off to freeze changes before restores begin back in the production environment. These restores typically utilize Veeam Instant Recovery where possible, to minimize downtime during the restore process. Alternatively, per-VM blackout windows can be utilized with further incremental backups to perform full VM or bare metal restores one at a time.