Github user StephanEwen commented on the pull request:
https://github.com/apache/flink/pull/202#issuecomment-90903241
What does the timeout measure? The time while no data was coming? Can the
protocol send heartbeats if no data is available?
Would it be simpler to switch to TCP as the coordinator protocol? It has
keepalive messages integrated and it throws an error on the reader side if the
connection is closed by the sender. Would that cover it?
A shutdown hook is a good idea. Make sure you also remove the shutdown hook
in `close()`, otherwise you have a resource leak in the JVM. Take a look at the
`BlobServer` and its `shutdown()` method for an example of how to do this
robustly.
---
If your project is set up for it, you can reply to this email and have your
reply appear on GitHub as well. If your project does not have this feature
enabled and wishes so, or if the feature is enabled but not working, please
contact infrastructure at infrastruct...@apache.org or file a JIRA ticket
with INFRA.
---
