Marie Hoeger created BEAM-12580:
-----------------------------------
Summary: Add ability to customize
Key: BEAM-12580
URL: https://issues.apache.org/jira/browse/BEAM-12580
Project: Beam
Issue Type: Improvement
Components: runner-universal
Affects Versions: 2.30.0
Environment: Python Linux
Reporter: Marie Hoeger
My development setup is all Docker-based and I work inside a Docker container
for development. From this docker environment, i'd love to be able to [run a
pipeline with a custom container
image|https://beam.apache.org/documentation/runtime/environments/] for local
testing. I've been able to launch a sister beam docker container from my main
docker container, but am running into an issue with the ports.
It would be really helpful if I could set a value for the control_address and
logging_endpoint of the GrpcServer for the portable runner worker.
The 'docker run' command passes in a control_address
[here|https://github.com/apache/beam/blob/2aab1b04c2f5002527e0f2d25075b282feb7c054/sdks/python/apache_beam/runners/portability/fn_api_runner/worker_handlers.py#L767-L769]
which ultimately comes from
[here.|https://github.com/apache/beam/blob/2aab1b04c2f5002527e0f2d25075b282feb7c054/sdks/python/apache_beam/runners/portability/fn_api_runner/worker_handlers.py#L464]
Similarly, the logging_endpoint passed in
[here|https://github.com/apache/beam/blob/2aab1b04c2f5002527e0f2d25075b282feb7c054/sdks/python/apache_beam/runners/portability/fn_api_runner/worker_handlers.py#L766]
comes from
[here|https://github.com/apache/beam/blob/2aab1b04c2f5002527e0f2d25075b282feb7c054/sdks/python/apache_beam/runners/portability/fn_api_runner/worker_handlers.py#L515].
This would be a pretty small change, just exposing some knob to configure those
endpoints rather than having them chosen by the gRPC server. Thanks!!
--
This message was sent by Atlassian Jira
(v8.3.4#803005)
