How to fix setting "indices.recovery.max_bytes_per_sec" in splitted cluster?

bump