I've created a Trident topology with persistentAggregate, where I save aggregated data to a DB. I've created a fake BatchSpout for my tests. I noticed that when running in local mode, every time I start my topology, the txid values are reset, which in practice resets the counters that I stored in my previous run.
My intention in production is to use a Trident Kafka spout. I'm trying to figure out what will happen to my real data when I run the topology. Will restarting the topology cause the spout to reset the txid value, which will make me end up with data loss? What will happen if I switch to a different Kafka/Zookeeper cluster?