Re: Testing recoverable job state

2017-09-15 Thread Robert Metzger
Hi, 1. Question: When you are throwing an exception within your user code, Flink will cancel the execution of all tasks and schedule them again (if you've configured a restart strategy). 2. Question: You'll need to configure the MiniCluster in HA mode. I believe that should be possible by passing

Testing recoverable job state

2017-09-13 Thread Marchant, Hayden
I'm a newbie to Flink and am trying to understand how the recovery works using state backends. I've read the documentation and am now trying to run a simple test to demonstrate the abilities - I'd like to test the recovery of a flink job and how the state is recovered from where it left off when