Great, you figured it out. Thanks for posting this back to the list.
On Sat, Feb 24, 2024 at 5:23 PM George Dekermenjian <ged1...@gmail.com>
wrote:
> Adding the following to both Dockerfile.launcher and Dockerfile.worker did
> the trick for me.
>
>
>
> COPY --from=apache/beam_java11_sdk:latest /opt/apache/beam/jars
> /opt/apache/beam/jars COPY --from=apache/beam_java11_sdk:latest
> /opt/java/openjdk /opt/java/openjdk ENV JAVA_HOME=/opt/java/openjdk ENV
> PATH="${JAVA_HOME}/bin:${PATH}"
>
>
>
> On Sat, Feb 24, 2024 at 21:55 XQ Hu via user <user@beam.apache.org> wrote:
>
>> Does your code work without the launcher? Better check this step by step
>> to figure out which part causes this error.
>>
>> On Sat, Feb 24, 2024 at 3:25 AM George Dekermenjian <ged1...@gmail.com>
>> wrote:
>>
>>> I have a python pipeline that uses the bigquery storage write method
>>> (cross language with Java). I’m building launcher and worker docker images
>>> and then launching the flex template. The launcher fails due to the
>>> following runtime error in dataflow.
>>>
>>> I’m using runner v2 and it is a streaming pipeline using the streaming
>>> engine.
>>>
>>> Any ideas of what is causing this?
>>>
>>> RuntimeError:
>>> org.apache.beam.vendor.guava.v32_1_2_jre.com.google.common.util.concurrent.UncheckedExecutionException:
>>> org.apache.beam.vendor.guava.v32_1_2_jre.com.google.common.util.concurrent.UncheckedExecutionException:
>>> java.lang.ClassCastException: class
>>> org.apache.beam.sdk.transforms.windowing.Repeatedly cannot be cast to class
>>> org.apache.beam.sdk.transforms.windowing.Trigger$OnceTrigger
>>> (org.apache.beam.sdk.transforms.windowing.Repeatedly and
>>> org.apache.beam.sdk.transforms.windowing.Trigger$OnceTrigger are in unnamed
>>> module of loader 'app')"
>>> line: "exec.go:66"
>>>
>>>
