Create

Sign in to ReadmeX

or

How we handle your data: Privacy Policy

ReadmeX

Your AI reading companion

What would you like to catch up on?

Pick a briefing, or ask your own question.

Your community and people briefings stay personal to you.

Story

Ray Tutorial Appendix Covers CI/CD, Slurm, and Cloud Scheduling

AI summary

A Ray tutorial appendix explains CI/CD for Ray applications, deployment on Slurm clusters without Kubernetes, and cloud scheduling costs. It also covers integration with orchestrators such as Airflow and Prefect, including job submission, status polling, and retry behavior. The appendix marks unspecified versions, field names, and default values as unconfirmed and presents the example code as a framework rather than a ready-to-copy setup.

In-depth insight

The appendix gives Ray practitioners concrete guidance on testing and deploying distributed workloads across Kubernetes, Slurm, and cloud environments while clarifying when to use external orchestrators.

仓库地址:github.com/hhk-png/cyc…

这份附录回答三个正文没展开的问题:Ray 应用的 CI/CD 长什么样、 只有 Slurm 没有 Kubernetes 的集群上怎么跑 Ray、 云上这套东西到底花多少钱。最后补一节"Ray 和 Airflow 这类编排器怎么接"。 四节按同一个标准写:能给全的命令给全,给不出的版本号、字段名、默认值 一律写 未确认 —— 示例代码请当骨架读,不要当"照抄就能跑"。


G.1 本章定位

第 17 章讲了部署形态,但对比表里缺了一列:代码是怎么上去的。 第 19 章把"集群与资源"画成第 1 层,却只列了 Kubernetes / Slurm / 云 VM 三个名字。 这份附录把这两处补上。

"Ray 在我这儿到底跑在哪"只有三个真实答案:

#Ray#KubeRay Operator#Kubernetes#Slurm#Airflow#Prefect#Anyscale#pytest#Ray Serve#Ray Jobs

Original · 39 min readRead the original
Group chat

No comments yet. Start the conversation.