Details
-
New Feature
-
Status: Resolved
-
P2
-
Resolution: Won't Fix
-
None
Description
Currently, blockOnRun is implemented in different ways by runners.
DirectRunner did blockOnRun based on DirectOptions.isBlockOnRun.
Dataflow have a separate BlockingDataflowRunner.
Flink and Spark runners might or might not block depends on their implementation on run().
I think DirectRunner's approach is the right way to go, and isBlockOnRun options need to be promoted to the general PipelineOptions.
Attachments
Issue Links
- links to