Hi, I am trying to use kuberay operator and ray cluster with in the Flyte by following this documentation. https://blog.flyte.org/ray-and-flyte#heading-launch-a-ray-cluster I have deployed the kuberay-operator and ray-cluster by following the helm chart docs given here https://github.com/ray-project/kuberay/tree/master/helm-chart. The pods are up and running. I have also enabled the ray plugin in the flyte. When running the pyflyte remotely, I am getting this below error.(I tried deploying the ray pods both in separate namespace as well as in the namespace where flyte pods are running). Need help here on this error.

raise _InactiveRpcError(state)
grpc._channel._InactiveRpcError: <_InactiveRpcError of RPC that terminated with:
        status = StatusCode.INTERNAL
        details = "failed to create workflow in propeller json: error calling MarshalJSON for type *v1alpha1.Inputs: Marshal called with nil"
        debug_error_string = "{"created":"@1666318439.715331527","description":"Error received from peer ipv4:35.172.9.181:443","file":"src/core/lib/surface/call.cc","file_line":1074,"grpc_message":"failed to create workflow in propeller json: error calling MarshalJSON for type *v1alpha1.Inputs: Marshal called with nil","grpc_status":13}"

Hi @delightful-lion-57360 this doesn't seem to be a Ray specific error. Rather, this indicates that FlyteAdmin is unable to create the FlyteWorkflow CRD due to the specified error. Can you provide the workflow definiton and the pyflyte command you are using? It seems for some reason there is an error in marshalling input values.

Thank you. Below is the code and command. This code is referred from the blog that i have pinged above. Code:

import ray
import os
os.system("pip3 install flytekitplugins-ray")
from typing import *
import typing
from flytekit import task
from flytekitplugins.ray import HeadNodeConfig, RayJobConfig, WorkerNodeConfig

@ray.remote
def f(x):
    return x * x


@task(task_config=RayJobConfig(head_node_config=HeadNodeConfig(ray_start_params={"block": "true"}),worker_node_config=[WorkerNodeConfig(group_name="ray-group", replicas=5, min_replicas=2, max_replicas=10)],runtime_env={"pip": ["numpy", "pandas"]}, ))
def ray_task() -> typing.List[int]:
    futures = [f.remote(i) for i in range(10)]
    return ray.get(futures)

Command:

pyflyte run --remote kuberay_test.py ray_task

Can you try defining a workflow that calls the

ray_task

task? I do not believe

pyflyte run

supports executing tasks. You should be able to do something as simple as:

@workflow
def ray_wf():
  ray_task()

and then attempt to execute the workflow with

pyflyte run --remote kuberay_test.py ray_wf

It still gives the same error 😞

What’s your flytekit version, @delightful-lion-57360?

My flytekit version is: 1.1.1

Can you upgrade to 1.2.1 and let me know if the same error’s cropping up?

Yeah, I am getting same error still...

After I added Works for me after I added an input to `ray_task`:

import ray

import typing
from flytekit import task
from flytekitplugins.ray import HeadNodeConfig, RayJobConfig, WorkerNodeConfig


@ray.remote
def f(x):
    return x * x


@task(
    task_config=RayJobConfig(
        head_node_config=HeadNodeConfig(ray_start_params={"block": "True"}),
        worker_node_config=[
            WorkerNodeConfig(
                group_name="ray-group", replicas=5, min_replicas=2, max_replicas=10
            )
        ],
        runtime_env={"pip": ["numpy", "pandas"]},
    )
)
def ray_task(n: int) -> typing.List[int]:
    futures = [f.remote(i) for i in range(n)]
    return ray.get(futures)

Indeed, @tall-lock-23197. In fact, as @hallowed-mouse-14616 mentioned, this is not specific to Ray. I opened https://github.com/flyteorg/flyte/issues/3008 to track this bug. In the meantime, @tall-lock-23197, can we amend the blogpost to include a parameter and also use

typing.List

as the return type for the ray task?

The workflow is able to be registered with the flyte. Thank you for the help.