Hi All,
I'm looking for a way to distinguish between normal and abnormal
termination of a spark streaming application with (checkpointing enabled).
Adding application listener doesn't really help because onApplicationEnd
event has no information regarding the cause of the termination.
ssc_.sc.addSparkListener(new SparkListener {
override def onApplicationEnd(applicationEnd:
SparkListenerApplicationEnd): Unit = {
I need to manage an internal metadata if streaming application has been
terminated and that termination is not recoverable I have to delete the
metadata (state of stream in this particular application).
--
Sincerely,
Igor Makhlin
