[PR #7] [MERGED] Fix Mooncake EPD integration proxy flag #6

New issue

Closed

opened 2026-10-29 21:06:04 +00:00 by bsakaguchi · 0 comments

bsakaguchi commented

2026-10-29 21:06:04 +00:00

Owner

📋 Pull Request Information

Original PR: https://github.com/kvcache-ai/vllm/pull/7
Author: @jiangkuaixue123
Created: 8/29/2026
Status: ✅ Merged
Merged: 8/29/2026
Merged by: @gty111

Base: add-mooncake-ec-connector ← Head: codex/fix-mooncake-epd-proxy-flag

📝 Commits (1)

8789a18 Fix Mooncake EPD proxy flag in integration script

📊 Changes

1 file changed (+1 additions, -1 deletions)

View changed files

📝 tests/v1/ec_connector/integration/run_epd_mooncake_ec_full_pipeline.sh (+1 -1)

📄 Description

Summary

update the Mooncake EPD full-pipeline script to use the proxy current --ec-consumer-zmq-addrs option
prevent argparse from rejecting the integration proxy launch

Duplicate-work check

This targets the feature branch behind vllm-project/vllm#41567. I searched open PRs in both vllm-project/vllm and kvcache-ai/vllm for this flag mismatch and found no existing fix.

Tests

git diff --check
bash -n tests/v1/ec_connector/integration/run_epd_mooncake_ec_full_pipeline.sh
verified the old option no longer appears in the proxy/script execution path and the script matches the parser option

The GPU full-pipeline test was not run because this change was prepared without the required GPU/Mooncake runtime.

AI assistance

AI assistance was used to identify the stale option, prepare the one-line fix, and run the static checks. The human submitter should review and understand the changed line before upstreaming.

_{🔄 This issue represents a GitHub Pull Request. It cannot be merged through Gitea due to API limitations.}

## 📋 Pull Request Information **Original PR:** https://github.com/kvcache-ai/vllm/pull/7 **Author:** [@jiangkuaixue123](https://github.com/jiangkuaixue123) **Created:** 8/29/2026 **Status:** ✅ Merged **Merged:** 8/29/2026 **Merged by:** [@gty111](https://github.com/gty111) **Base:** `add-mooncake-ec-connector` ← **Head:** `codex/fix-mooncake-epd-proxy-flag` --- ### 📝 Commits (1) - [`8789a18`](https://github.com/kvcache-ai/vllm/commit/8789a1886cc7783bdd2f19bc8918a0985f97d0e7) Fix Mooncake EPD proxy flag in integration script ### 📊 Changes **1 file changed** (+1 additions, -1 deletions) <details> <summary>View changed files</summary> 📝 `tests/v1/ec_connector/integration/run_epd_mooncake_ec_full_pipeline.sh` (+1 -1) </details> ### 📄 Description ## Summary - update the Mooncake EPD full-pipeline script to use the proxy current `--ec-consumer-zmq-addrs` option - prevent argparse from rejecting the integration proxy launch ## Duplicate-work check This targets the feature branch behind vllm-project/vllm#41567. I searched open PRs in both vllm-project/vllm and kvcache-ai/vllm for this flag mismatch and found no existing fix. ## Tests - `git diff --check` - `bash -n tests/v1/ec_connector/integration/run_epd_mooncake_ec_full_pipeline.sh` - verified the old option no longer appears in the proxy/script execution path and the script matches the parser option The GPU full-pipeline test was not run because this change was prepared without the required GPU/Mooncake runtime. ## AI assistance AI assistance was used to identify the stale option, prepare the one-line fix, and run the static checks. The human submitter should review and understand the changed line before upstreaming. --- <sub>🔄 This issue represents a GitHub Pull Request. It cannot be merged through Gitea due to API limitations.</sub>

bsakaguchi

2026-10-29 21:06:04 +00:00