Hey folks. I have Flink setup on a single server where I keep losing the task manager and having to restart it. It always happens at the end of executing a batch job. I have multiple streaming jobs and only one batch job. I have been trying to debug to no avail for hours. Have tried setting the following parameters but none seem to help:

akka.ask.timeout
akka.lookup.timeout
akka.tcp.timeout
heartbeat.interval
heartbeat.rpc-failure-threshold
heartbeat.timeout