In 2.52, we're introducing ray symmetric-run -- a CLI command to simplify and improve the large model interactive development experience on Ray and @vllm_project!
Spinning up multi-node vLLM with Ray on interactive environments can be tedious, requiring users to juggle separate nodes, breaking the “single entrypoint” mental model that many users expect.
Ray now has a new command: ray symmetric-run. This command makes it possible to launch the same entrypoint command on every node in a Ray cluster -- simplifying the process of spawning vLLM servers with multi-node models on HPC setups or in using parallel ssh tools like mpssh.
Check out the blog: blog.vllm.ai/2025/11/22/ray…
Spinning up multi-node vLLM with Ray on interactive environments can be tedious, requiring users to juggle separate nodes, breaking the “single entrypoint” mental model that many users expect.
Ray now has a new command: ray symmetric-run. This command makes it possible to launch the same entrypoint command on every node in a Ray cluster -- simplifying the process of spawning vLLM servers with multi-node models on HPC setups or in using parallel ssh tools like mpssh.
Check out the blog: blog.vllm.ai/2025/11/22/ray…
Nov 26, 2025 · 10:31 PM UTC
1
15
43
20,919
@raydistributed @vllm_project Sounds great as I’ve been struggling to get it working on my two DGX Sparks. Question, using Docker image of vllm on both Sparks head and worker, should I run the symmetric-run cmd inside each container?
1
0
0
156