Remote "Pull" Replication resync after site failure

[ccoope@cmers.com]

Looking into a design issue. We are looking to use Veeam to Replicate VM's to a remote site from a primary datacenter. The primary datacenter has a second server room with SAN-SAN replication enabled, where we can stand the entire environment up quickly if necessary, and be running again. The problem is that the second server room is running different hosts, which would be added into the original vCenter during the recovery stage along with all the other VM's. The Volumes would mount with the same names and everything, and it's the same vCenter, but I assume the registrations to the new hosts would change the VM ID's, and blow up the mappings in Veeam Replication jobs. Would the remote replication be able to pick back up again, or would it require a new initial sync to get restarted?

[Alexey D.]

Hello Chris,

In B&R console, how did you specify the source for your replication jobs, by adding vCenter or by standalone hosts? If by vCenter, yes, the mappings will be blown up because Vmware generates new VM IDs each time a new host is presented.

If by standalone hosts, there is a chance the mappings will remain. If your SAN-to-SAN job includes replication of ESX hypervisor images as well, probably it could be possible for new hosts to pick up those images and thus keep old IDs. Then even if hosts would be added to vCenter, B&R would still be able to communicate to them (referring by old IDs).

Just an idea, but maybe it makes sense.