venice

As long as you are careful to treat the data directory as one atomic unit, you can do pretty much whatever you want with it… if you lose it, it’ll rebootstrap. If you copy it from elsewhere, it’ll resume from the checkpoint where the other instance left off (assuming that Helix re-assigns the same partitions to this node, which normally it would do if the hostname is the same, but in some scenarios it might not, and then it would rebootstrap).

I say “one atomic unit” because there is some part of the data directory where the progress gets checkpointed. If you were to keep that and delete the data, then it won’t rebootstrap, and you’ll effectively have a data gap.