Non-computing types use ray clusters

How severe does this issue affect your experience of using Ray?

  • High: It blocks me to complete my task.

I’m using Microsoft AZURE and I’m going to implement my work tasks with the help of ray. My work tasks are not computational but they also need multi-process (multi-CPU) support, but I’m very confused by a few questions.

  1. if I use the example-full.yaml method to generate an auto-scaling cluster on AZURE, do I need to manually configure the python runtime environment for each node? I know that there are commands that can be added to the yaml file to initialise the node environment, but my business needs the same runtime environment, specifically downloading the chrome browser for each node, then downloading the same version of the chromedriver driver, then putting the driver in the specified folder, in addition to the python 3.10 installation, and the I don’t know if it is suitable to deploy my business stream using example-full.yaml.

I don’t understand how ray clustering is implemented on AZURE, how much time it takes to scale a node, and how every second is needed for my business, is it better for my business to build ray clusters manually.

I’m still confused about how to submit tasks to the ray cluster, I’ve read about the commands in ray submit, it would be great to have a sample code in a production environment, I’m a noob but I love the ray project.

Hi all, I have learned clearly how to submit tasks to ray cluster through ray documentation and help from ray team.