Can be jobs allocated to different OS machines?

Ray Core
1

Dear all

I am introducing ray for distributed computing.
As my questions,
can be jobs allocated to different OS machines,
that is, the head is Windows, then can be job submitted to Linux OS machines?
(If you know details, I hope that you tell me limits, such as file I/O is prohibited and so on)

Thank you in advance.

regards

1. Severity of the issue: (select one)
Medium: Significantly affects my productivity but can find a workaround.

2. Environment:

  • Ray version: the latest
  • Python version: 3.12
  • OS: Windows 11 (+ RedHat)
  • Cloud/Infrastructure: Nan
  • Other libs/tools (if relevant):

3. What happened vs. what you expected:

  • Expected:
  • Actual:
2

Hi mit - welcome to the Ray community!

Yes, Jobs can be allocated to different OS machines e.g. the head node can be a Windows machine and Jobs can be submitted to Linux OS machines. Keep in mind there are some limitations when using Windows, like multi-node Ray clusters being untested on Windows. Also, Jobs are bound to the lifetime of a Ray cluster, so if the cluster goes down, all running Jobs on that cluster will be terminated. Sharing this link for reference: Environment Dependencies — Ray 2.44.0. Please let me know if this helps!