Thank, looks like this bugs,
How workaround I tried set withWatermarkFn2 in Instant.now()
(like NowTimestampFn in KafkaIO)
DirectRunner - nothing changed
FlinkRunner - trigger started work correct
Can this behavior be caused by UnboundedReadEvaluatorFactory line 142?
Some description of my pipeline
kafka -> transform1 -> window.into (10s) -> groupBy -> output1
|-> transform2 -> output2
I alway can see output2
but output1 work unstable
On 2016-11-07 10:07 (+0400), Raghu Angadi <r...@google.com.INVALID> wrote:
> On Sun, Nov 6, 2016 at 4:31 AM, Demin Alexey <di...@gmail.com> wrote:>
>
> > This is bug or incorrect using API from my side?>
>
>
> This is a bug in KafkaIO. It should advance the watermark when there are
no>
> messages to read. https://issues.apache.org/jira/browse/BEAM-591>
>
> I want to fix it, but may not get to it until Thanksgiving, will see.>
>
