Hi, I have a question regarding kafka connect and how it operates in distributed mode. We use docker swarm to deploy kafka connect and only one instance of it in swarm. After node crashing with kafka connect and starting another instance on a new node it starts to read configuration topic from offset 385 despite seeking to beginning of the topic. As I checked in kafka connect code in KafkaBasedLog class it always tries to read configuration topic from the beginning even if there are some committed offsets, however in our case it reads from the last offset and no configuration of connectors is read. Does anyone have any idea why it happens and how it could be fixed?
Thanks.
