cozos opened a new issue, #26756:
URL: https://github.com/apache/beam/issues/26756
### What happened?
In Beam on Dataflow, we are getting Python SDK Harness failures that
manifest in the job logs like this:
```
Error message from worker: SDK harness sdk-0-1 disconnected. This usually
means that the process running the pipeline code has crashed. Inspect the
Worker Logs and the Diagnostics tab to determine the cause of the crash.
```
Looking through the logs, the root cause for the failure is dropping the
gRPC connection for Fn Control API:
```
Python sdk harness failed:
Traceback (most recent call last):
File
"/bag_data_cache_docker_pybinary.runfiles/cruise_ws/cruise/mlp/cfs/projects/trajectory_ranking/pipeline/bag_data_cache_docker_pybinary_exedir/apache_beam/runners/worker/sdk_worker_main.py",
line 157, in main
sdk_harness.run()
File
"/bag_data_cache_docker_pybinary.runfiles/cruise_ws/cruise/mlp/cfs/projects/trajectory_ranking/pipeline/bag_data_cache_docker_pybinary_exedir/apache_beam/runners/worker/sdk_worker.py",
line 234, in run
for work_request in self._control_stub.Control(get_responses()):
File
"/bag_data_cache_docker_pybinary.runfiles/cruise_ws/cruise/mlp/cfs/projects/trajectory_ranking/pipeline/bag_data_cache_docker_pybinary_exedir/grpc/_channel.py",
line 426, in __next__
return self._next()
File
"/bag_data_cache_docker_pybinary.runfiles/cruise_ws/cruise/mlp/cfs/projects/trajectory_ranking/pipeline/bag_data_cache_docker_pybinary_exedir/grpc/_channel.py",
line 809, in _next
raise self
grpc._channel._MultiThreadedRendezvous: <_MultiThreadedRendezvous of RPC
that terminated with:
status = StatusCode.UNAVAILABLE
details = "keepalive watchdog timeout"
debug_error_string =
"{"created":"@1684279532.354148410","description":"Error received from peer
ipv6:[::1]:12371","file":"src/core/lib/surface/call.cc","file_line":1069,"grpc_message":"keepalive
watchdog timeout","grpc_status":14}"
```
Ignoring the root cause for now (although if some Dataflow experts can help
shed some light on this, it'd be appreciated) - it was pretty hard to actually
find that `Python sdk harness failed` log message.
This is because the log severity of this stack trace is just `ERROR` - but I
would expect a problem that brought down a worker process to be `CRITICAL`!
It's important that we understand what killed our SDK Harness.
### Issue Priority
Priority: 2 (default / most bugs should be filed as P2)
### Issue Components
- [X] Component: Python SDK
- [ ] Component: Java SDK
- [ ] Component: Go SDK
- [ ] Component: Typescript SDK
- [ ] Component: IO connector
- [ ] Component: Beam examples
- [ ] Component: Beam playground
- [ ] Component: Beam katas
- [ ] Component: Website
- [ ] Component: Spark Runner
- [ ] Component: Flink Runner
- [ ] Component: Samza Runner
- [ ] Component: Twister2 Runner
- [ ] Component: Hazelcast Jet Runner
- [ ] Component: Google Cloud Dataflow Runner
--
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.
To unsubscribe, e-mail: [email protected]
For queries about this service, please contact Infrastructure at:
[email protected]
