Enhance run_cluster.sh for multi-NIC support

[evberrypi]

Purpose

Added support for multi-NIC configurations by parsing VLLM_HOST_IP from additional arguments. Built Ray IP environment variables to ensure correct network interface binding.

Test Plan

Run on device with multiple interfaces (such as DGX Spark) . The desired behavior is that the IP address of the NICs can be specified (previously they default to the IP of the host)
bash run_cluster.sh $VLLM_IMAGE 192.168.100.10 --head ~/.cache/huggingface
-e VLLM_HOST_IP=192.168.100.10
-e UCX_NET_DEVICES=$MN_IF_NAME
-e NCCL_SOCKET_IFNAME=$MN_IF_NAME
-e OMPI_MCA_btl_tcp_if_include=$MN_IF_NAME
-e GLOO_SOCKET_IFNAME=$MN_IF_NAME
-e TP_SOCKET_IFNAME=$MN_IF_NAME
-e RAY_memory_monitor_refresh_ms=0
-e MASTER_ADDR=192.168.100.10

Test Result

---

Essential Elements of an Effective PR Description Checklist

• The purpose of the PR, such as "Fix some issue (link existing issues this PR will resolve)".
• The test plan, such as providing test command.
• [] The test results, such as pasting the results comparison before and after, or e2e results
• (Optional) The necessary documentation update, such as updating supported_models.md and examples for a new model.
• (Optional) Release notes update. If your change is user facing, please update the release notes draft in the Google Doc.

[github-actions]

👋 Hi! Thank you for contributing to the vLLM project.

💬 Join our developer Slack at https://slack.vllm.ai to discuss your PR in #pr-reviews, coordinate on features in #feat- channels, or join special interest groups in #sig- channels.

Just a reminder: PRs would not trigger full CI run by default. Instead, it would only run fastcheck CI which starts running only a small and essential subset of CI tests to quickly catch errors.

You ask your reviewers to trigger select CI tests on top of fastcheck CI.

Once the PR is approved and ready to go, your PR reviewer(s) can run CI to test the changes comprehensively before merging.

To run CI, PR reviewers can either: Add ready label to the PR or enable auto-merge.

If you have any questions, please reach out to us on Slack at https://slack.vllm.ai.

🚀

[mergify]

Documentation preview: https://vllm--28328.org.readthedocs.build/en/28328/

[gemini-code-assist]

Code Review

This pull request enhances run_cluster.sh to support multi-NIC configurations by allowing VLLM_HOST_IP to be specified. The core change involves parsing this IP from the command-line arguments. My review identifies a potential bug in the parsing logic which could lead to incorrect behavior if arguments are passed in an unexpected order. I've provided a more robust implementation to address this.

[gemini-code-assist]

The current logic for parsing VLLM_HOST_IP is not robust. It will find and use a VLLM_HOST_IP=... argument even if it is not preceded by a -e flag, which is required for passing environment variables to docker run. This could lead to misinterpreting arguments intended for the command inside the container. The parsing should be stricter to only identify VLLM_HOST_IP when it's explicitly passed as an environment variable.

Suggested change

	VLLM_HOST_IP=""
	for arg in "${ADDITIONAL_ARGS[@]}"; do
	if [[ $arg == "-e" ]]; then
	continue
	fi
	if [[ $arg == VLLM_HOST_IP=* ]]; then
	VLLM_HOST_IP="${arg#VLLM_HOST_IP=}"
	break
	fi
	done
	VLLM_HOST_IP=""
	prev_arg=""
	for arg in "${ADDITIONAL_ARGS[@]}"; do
	if [[ "${prev_arg}" == "-e" ]] && [[ "${arg}" == VLLM_HOST_IP=* ]]; then
	VLLM_HOST_IP="${arg#VLLM_HOST_IP=}"
	break
	fi
	prev_arg="${arg}"
	done