Shared connection to head node closed

Ray Clusters
1

1. Severity of the issue: (select one)
High: Completely blocks me.

2. Environment:

  • Ray version: 2.10
  • Python version: 3.8
  • OS: Rhel 8
  • Cloud/Infrastructure: In premises
  • Other libs/tools (if relevant):

3. What happened vs. what you expected:

  • Expected: On ray dashboard workers node should be listed
  • Actual: only head node is avaible on cluster page

I defined worker’s ip in yaml file and min and max workers are set to2 , when i hit ray up cluster.yaml command only single error come Shared connection to head ip closed and in ray dashboard no active node visible

2

If only the head node appears in the Ray dashboard and no worker nodes are listed, even though you defined worker IPs and set min/max workers, this usually means Ray is not starting or connecting on the worker nodes. In local provider mode, Ray does not automatically start workers; you must manually SSH into each worker and run ray start --address=‘head_ip:6379’ on each worker node. The “Shared connection to head ip closed” message is normal and not an error by itself. Also, ensure SSH access and network connectivity between nodes are working, and that all required ports are open. See details in this discussion.

Would you like a step-by-step guide for manual worker startup and troubleshooting?

Sources:

Hint: Mention @RunLLM in the post for followups.

3

Thanks for let me know about local provider