Hi, I have a question regarding running flink locally: does local source files (flink-conf.yaml, plugins) gets reloaded when one submit a new flink job, or one has to restart the cluster in order to do that? (Also I noticed that even after one uses

stop-cluster.sh

there are still some task-manager dangling in

ps aux

(presumably keeping the job history). Does one needs to kill these in order to help reload the flink source files? Does that leads to memory leaks if too many old task-manager processes hanging around?

Changes to plugins or flink-conf.yaml only take affect upon a restart of the process (TM, JM). I think_,_ stop-cluster.sh only stops one taskmanager. Not sure if this is a bug or intended.

taskmanager.sh stop-all

can be used to stop all taskmanagers.

I think_,_ stop-cluster.sh only stops one taskmanager

That depends. If you manually start additional task managers then they will not be stopped. If you setup all processes via

conf/workers

then it will stop all of them.

Thanks for the response. Currently, I am just greping all flink processes after the

stop-cluster.sh

and kill them all.