aaronrobert·7 mesi fa·discussIt seems that the data is written to only a single node. But if that's the case, how is fault-tolerant ensured? In terms of data synchronization, how is this different from setting the replication factor to one in Kafka or Pulsar?