Example #1
0
def grpc_schedule_origin(schedule_name):
    loadable_target_origin = LoadableTargetOrigin(
        executable_path=sys.executable, python_file=__file__, attribute="the_repo"
    )
    server_process = GrpcServerProcess(loadable_target_origin=loadable_target_origin)
    with server_process.create_ephemeral_client() as api_client:
        repo_origin = RepositoryGrpcServerOrigin(
            host=api_client.host,
            port=api_client.port,
            socket=api_client.socket,
            repository_name="the_repo",
        )
        yield repo_origin.get_schedule_origin(schedule_name)
    server_process.wait()
Example #2
0
 def get_origin(self):
     if isinstance(self.repository_location_handle,
                   InProcessRepositoryLocationHandle):
         return RepositoryPythonOrigin(
             code_pointer=self.repository_location_handle.
             repository_code_pointer_dict[self.repository_name],
             executable_path=sys.executable,
         )
     elif isinstance(self.repository_location_handle,
                     ManagedGrpcPythonEnvRepositoryLocationHandle):
         return RepositoryPythonOrigin(
             code_pointer=self.repository_location_handle.
             repository_code_pointer_dict[self.repository_name],
             executable_path=self.repository_location_handle.
             executable_path,
         )
     elif isinstance(self.repository_location_handle,
                     GrpcServerRepositoryLocationHandle):
         return RepositoryGrpcServerOrigin(
             host=self.repository_location_handle.host,
             port=self.repository_location_handle.port,
             socket=self.repository_location_handle.socket,
             repository_name=self.repository_name,
         )
     else:
         check.failed(
             "Can not target represented RepositoryDefinition locally for repository from a {}."
             .format(self.repository_location_handle.__class__.__name__))
Example #3
0
def sync_get_external_repositories_grpc(api_client,
                                        repository_location_handle):
    check.inst_param(repository_location_handle, "repository_location_handle",
                     RepositoryLocationHandle)

    repos = []
    for repository_name in repository_location_handle.repository_names:
        external_repository_data = check.inst(
            api_client.external_repository(
                repository_grpc_server_origin=RepositoryGrpcServerOrigin(
                    repository_location_handle.host,
                    repository_location_handle.port,
                    repository_location_handle.socket,
                    repository_name,
                )),
            ExternalRepositoryData,
        )
        repos.append(
            ExternalRepository(
                external_repository_data,
                RepositoryHandle(
                    repository_name=external_repository_data.name,
                    repository_location_handle=repository_location_handle,
                ),
            ))
    return repos
def sync_get_streaming_external_repositories_grpc(api_client,
                                                  repository_location_handle):
    check.inst_param(repository_location_handle, "repository_location_handle",
                     RepositoryLocationHandle)

    repos = []
    for repository_name in repository_location_handle.repository_names:
        external_repository_chunks = list(
            api_client.streaming_external_repository(
                repository_grpc_server_origin=RepositoryGrpcServerOrigin(
                    repository_location_handle.host,
                    repository_location_handle.port,
                    repository_location_handle.socket,
                    repository_name,
                )))

        external_repository_data = deserialize_json_to_dagster_namedtuple(
            "".join([
                chunk["serialized_external_repository_chunk"]
                for chunk in external_repository_chunks
            ]))

        repos.append(
            ExternalRepository(
                external_repository_data,
                RepositoryHandle(
                    repository_name=external_repository_data.name,
                    repository_location_handle=repository_location_handle,
                ),
            ))
    return repos
Example #5
0
def get_repository_origin_from_kwargs(kwargs):
    load_target = created_workspace_load_target(kwargs)

    if isinstance(load_target, PythonFileTarget):
        return RepositoryPythonOrigin(
            executable_path=sys.executable,
            code_pointer=CodePointer.from_python_file(
                load_target.python_file,
                load_target.attribute,
                working_directory=load_target.working_directory,
            ),
        )
    elif isinstance(load_target, ModuleTarget):
        return RepositoryPythonOrigin(
            executable_path=sys.executable,
            code_pointer=CodePointer.from_module(load_target.module_name,
                                                 load_target.attribute),
        )
    elif isinstance(load_target, GrpcServerTarget):
        return RepositoryGrpcServerOrigin(
            host=load_target.host,
            port=load_target.port,
            socket=load_target.socket,
            repository_name=kwargs['repository'],
        )
    else:
        check.failed('invalid')
Example #6
0
def test_grpc_server_down():
    with _default_instance() as instance:
        down_grpc_repo_origin = RepositoryGrpcServerOrigin(
            host="localhost", port=find_free_port(), socket=None, repository_name="down_repo"
        )
        down_grpc_schedule_origin = down_grpc_repo_origin.get_schedule_origin("down_schedule")

        instance = DagsterInstance.get()
        result = sync_launch_scheduled_execution(down_grpc_schedule_origin)

        assert isinstance(result, ScheduledExecutionFailed)
        assert "failed to connect to all addresses" in result.errors[0].to_string()

        ticks = instance.get_schedule_ticks(down_grpc_schedule_origin.get_id())
        assert ticks[0].status == ScheduleTickStatus.FAILURE
        assert "failed to connect to all addresses" in ticks[0].error.message
def grpc_schedule_origin(schedule_name):
    with seven.TemporaryDirectory() as temp_dir:
        with environ({'DAGSTER_HOME': temp_dir}):
            loadable_target_origin = LoadableTargetOrigin(
                executable_path=sys.executable,
                python_file=__file__,
                attribute='the_repo')
            with GrpcServerProcess(
                    loadable_target_origin=loadable_target_origin
            ).create_ephemeral_client() as api_client:
                repo_origin = RepositoryGrpcServerOrigin(
                    host=api_client.host,
                    port=api_client.port,
                    socket=api_client.socket,
                    repository_name='the_repo',
                )
                yield repo_origin.get_schedule_origin(schedule_name)
def test_cancel_run():
    with instance_for_test() as instance:

        loadable_target_origin = LoadableTargetOrigin(
            executable_path=sys.executable, python_file=__file__, working_directory=None,
        )

        server_process = GrpcServerProcess(loadable_target_origin, max_workers=10)

        with server_process.create_ephemeral_client() as api_client:
            streaming_results = []

            pipeline_run = instance.create_run_for_pipeline(
                streaming_pipeline, run_config={"solids": {"streamer": {"config": {"length": 20}}}},
            )
            execute_run_args = ExecuteRunArgs(
                pipeline_origin=PipelineGrpcServerOrigin(
                    pipeline_name="streaming_pipeline",
                    repository_origin=RepositoryGrpcServerOrigin(
                        host="localhost",
                        socket=api_client.socket,
                        port=api_client.port,
                        repository_name="test_repository",
                    ),
                ),
                pipeline_run_id=pipeline_run.run_id,
                instance_ref=instance.get_ref(),
            )
            stream_events_result_thread = threading.Thread(
                target=_stream_events_target, args=[streaming_results, api_client, execute_run_args]
            )
            stream_events_result_thread.daemon = True
            stream_events_result_thread.start()
            poll_for_step_start(instance, pipeline_run.run_id)

            res = api_client.cancel_execution(
                cancel_execution_request=CancelExecutionRequest(run_id=pipeline_run.run_id)
            )
            assert res.success is True

            poll_for_run(instance, pipeline_run.run_id)

            logs = instance.all_logs(pipeline_run.run_id)
            assert (
                len(
                    [
                        ev
                        for ev in logs
                        if ev.dagster_event.event_type_value == "STEP_MATERIALIZATION"
                    ]
                )
                < 20
            )

            # soft termination
            assert [ev for ev in logs if ev.dagster_event.event_type_value == "STEP_FAILURE"]

        server_process.wait()
Example #9
0
def get_repository_origin_from_kwargs(kwargs):
    load_target = created_workspace_load_target(kwargs)

    if isinstance(load_target, GrpcServerTarget):
        return RepositoryGrpcServerOrigin(
            host=load_target.host,
            port=load_target.port,
            socket=load_target.socket,
            repository_name=kwargs["repository"],
        )

    return get_repository_python_origin_from_kwargs(kwargs)
def test_grpc_server_down():
    with seven.TemporaryDirectory() as temp_dir:
        with environ({'DAGSTER_HOME': temp_dir}):
            down_grpc_repo_origin = RepositoryGrpcServerOrigin(
                host='localhost',
                port=find_free_port(),
                socket=None,
                repository_name='down_repo')
            down_grpc_schedule_origin = down_grpc_repo_origin.get_schedule_origin(
                'down_schedule')

            instance = DagsterInstance.get()
            result = sync_launch_scheduled_execution(down_grpc_schedule_origin)

            assert isinstance(result, ScheduledExecutionFailed)
            assert 'failed to connect to all addresses' in result.errors[
                0].to_string()

            ticks = instance.get_schedule_ticks(
                down_grpc_schedule_origin.get_id())
            assert ticks[0].status == ScheduleTickStatus.FAILURE
            assert 'failed to connect to all addresses' in ticks[
                0].error.message