How severe does this issue affect your experience of using Ray?
- High: It blocks me to complete my task.
I could not build a distributed application by a cluster without k8s，it can descripte by next
1、i build a new worker，it include a new model and different python env
2、i make this container running，and it is run by ray，but it seem that i could not add it into a running static ray cluster without k8s
it is a important things in some ai infra scenario！
hope to get a solution or a suggestion，thank you very much
actually，it seems to have two different solutions
1、Separating the dashboard from the head node of the cluster，that it can not care the monitor message where from，it just a dashboard for different serve/node
2、the ray serve may need to independent，that make it can need not to care the python env
My thoughts may be a little simple, please advise
Run Multiple Applications in Different Containers — Ray 2.35.0
i find it can build by podman can solve this problem，but i want to know how to implement it on the docker，and how can i add serve and worker dynamic
1 Like
but it seem work in this node，could not use in the cluster