Replacing cluster that is XDCR source

So here is the situation. I have two clusters (ClusterA, ClusterB). ClusterA performs one-way replication to ClusterB, but ClusterA is rather old and wanted to phase it out with ClusterC. So I have replicated the data to make ClusterC comparable to ClusterA. The ultimate goal here is I want ClusterC to be the source for XDCR to ClusterB and ClusterA out of the picture. But I am fairly sure if I just set up replication outright it will just duplicate all the information in the bucket. Is that correct?

What are my choices here? Should I flush the bucket in ClusterB and repopulate via XDCR that way ClusterC? I am not really finding my in the way of documents about this type of scenario.

Anybody have any ideas?

Hi tippy_top,
It is safe to setup a new replication from Cluster C -> Cluster B and delete the replication from Cluster A -> Cluster B. XDCR will sync the bucket on Cluster B, so you will not have 2 instances of the same doc, but rather the newer version of the doc from Cluster C. Hope that helps!