i'd like to report a "strange" use case i made of Replication function in Veeam.
Task: we have a CMS for an internet site (usage close to 100 millions page views per month) and we needed to upgrade its version.
We were asked to clone all the components: db+backend (approx 1.6 TB) and frontends (approx 1 TB) so the developer team could operate the upgrades on the replicas.
This was necessary to have the old environment up and running while upgrade, and to have an easy fallback in case of issues.
The rough part was that we needed to stop writers from writing stories during the clone process, so the cloned backend VMs had the exact same content of the old ones.
As you can easily guess, cloning would require many hours and such a long stop was not possible.
The solution we have followed is to use Veeam Replication to have replicas of the backend VMs (within the same cluster of source VMs), which were synced every 30 minutes.
This way, when ready to start the upgrade process, we told to stop writing stories, performed the last sync (which required only a few minutes) and we were ready to start.
All has worked fine as expected, both in one aborted case (upgrade failed so we kept old VMs in production) both in a success case (upgrade succeeded, so we put in productions the replicated 'n' upgraded VMs and powered down the old ones).
I cannot imagine how we could have made without Veeam!
