Kai-Hsun Chen comments

Results 324 comments of


                                            Kai-Hsun Chen

trafficstars

when actor_id_ is nil ,call ray.get_runtime_context().actor_name will cause raise

Would you mind rebasing with the master branch? Thanks!

when actor_id_ is nil ,call ray.get_runtime_context().actor_name will cause raise

> @kevin85421 not mind, how need i do Sync your branch with the upstream master branch.

[Docs][KubeRay] add example for distributed checkpointing with kuberay and gcsfuse

@aslonnie I will review this one. I have already had an offline sync with Andrew today.

[Dashboard | Core] Some Ray subprocesses exited unexpectedly: dashboard [exit code=-11]

cc @alanwguo

[Core] `ray.get_runtime_context().namespace` returns incorrect namespace in Actor Tasks

> `ray.get_runtime_context()` returns the job config for entire Ray job, not specific task/actor. Hmm. The RuntimeContext's constructor is initialized with a worker process. I believe that RuntimeContext is specific to...

KubeRay Deployment Failure with Large ServeZip File in Working_Dir

cc @GeneDer @fishbone

[Core] ray.remote raises ValueError when used on torch IterableDataset

> IIUC, this is a python 3.9 bug. Do you know if it's fixed in 3.10+? No, Python 3.10 still has the issue. I am not sure whether to call...

[Core] ray.remote raises ValueError when used on torch IterableDataset

The premerge fails on `test_classmethod_genericalias` and `test_placement_group_4.py`. I tried to reproduce locally with Python 3.8 and Python 3.10, but I can't. * Python 3.10 * Python 3.8

[Core] ray.remote raises ValueError when used on torch IterableDataset

Python 3.9.0 still can't reproduce the issue.

[Core] ray.remote raises ValueError when used on torch IterableDataset

I installed the Python 3.9 artifact on [the Buildkite CI](https://buildkite.com/ray-project/premerge/builds/19301#018daef5-dd55-417a-9bab-f2f6c273ca36), and then replaced the Ray Python scripts with my local ones. I still can't reproduce the CI error.