I know there was a closed thread discussing this already, but I still want to make a feature request for this.
The problem with having it in yml file is it'll stay with the new cluster forever unless I restart the new cluster once reindex is done. This is very intrusive. I don't really know when I need to reindex from the old cluster. We have tons of data and I'm not going to reindex unless I really need it. And the requirement might change.
Having whitelist in yml is not really offering security either. It's a configuration on the initiator.
Anybody can create a new cluster with the whitelist and reindex that way.
I really need an easy way to get data from old cluster and that's it. The need to restart cluster is too intrusive. Both clusters are within the same network. Is security really a concern here?
Please reconsider allowing reindex on the fly. Maybe change security to something like setting up a temporary secret in the old cluster and the new cluster needs this secret for reindex, etc.
Owner of old cluster can always remove the secret afterward. I believe this provides a much better security (which I don't believe is a concern in the first place), and also allow reindex on the fly.
Restarting cluster is not a good solution in general. If I have access to the new cluster, I can always query the old cluster.