nehsyc commented on a change in pull request #12678:
URL: https://github.com/apache/beam/pull/12678#discussion_r485991869
##########
File path:
runners/google-cloud-dataflow-java/src/main/java/org/apache/beam/runners/dataflow/DataflowPipelineTranslator.java
##########
@@ -1264,6 +1268,10 @@ private static void translateFn(
// in streaming but does not work in batch
if (context.getPipelineOptions().isStreaming() && isStateful) {
stepContext.addInput(PropertyNames.USES_KEYED_STATE, "true");
Review comment:
Update on this:
I added an experiment to gate the auto-sharding so this can be merged
without waiting for the backend. It will also make the testing easier.
I also added a check for the experiment, "beam_fn_api". My intention was to
disable the feature for unified worker but I guess this way we would disable
auto-sharding for both unified worker and java worker using fn api - I remember
that we are not going to support the latter so it seems fine to me. But let me
know if my understanding is incorrect.
----------------------------------------------------------------
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
For queries about this service, please contact Infrastructure at:
[email protected]
