Description
The janitor will only consider resetting the rebalance status when janitoring buckets (here), but this means that the rebalance status won't get reset if there are no buckets.
It's not clear what the original reason for only doing this when janitoring buckets was.
According to Ben Huddleston:
So I don't think it needs anything bucket related, but the position at which we do it is quite interesting. At this point we have connectivity to all other nodes, and this is in some leader activity, which is bit of a safety check against some rogue orchestrator flipping it when it should not
It could be separated, I believe, but it needs some thought
We should ensure that the rebalance status can be reset without any buckets, while taking care that we don't reset when we shouldn't.
Seen in CBSE-17605