I have two clusters, the primary is called A and the secondary is B.
I have indices in A for which I created follower indices in B.
Now, I could pause the following, close, unfollow and open the indices in B to make the follower indices as leaders.
I then tried to make the original leader indices on A as the followers of B, without deleting the indices.
I closed them, made them follow indices from B and opened them.
It did the trick, they turned into followers and I didn't lose data, they also synced up new documents + deletions from B, whatever happened when both were leader indices.
If so, then why does the Elasticsearch documentation and internet articles state that we need to DELETE the original indices from cluster A and re-create followers? Doesn't that mean losing so much data and time, while I have clusters with terabytes of data?