[Bug Fix] Fix address/port already in use error for pplx test #20094
Conversation
Signed-off-by: yewentao256 <zhyanwentao@126.com>
|
👋 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 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 🚀 |
There was a problem hiding this comment.
Summary of Changes
Hello @yewentao256, I'm Gemini Code Assist1! I'm currently reviewing this pull request and will post my feedback shortly. In the meantime, here's a summary to help you and other reviewers quickly get up to speed!
This pull request resolves a critical address already in use error that was causing intermittent failures in distributed training tests. By implementing dynamic port allocation for torch.distributed process group initialization, it ensures that tests can reliably find and use an available port, thereby enhancing the stability and consistency of the test suite.
Highlights
- Bug Fix - Port Contention: This pull request addresses the
EADDRINUSEerror (address already in use) that occurred duringtorch.distributedprocess group initialization, specifically when the default port 29500 was unavailable, leading to test failures. - Dynamic Port Allocation: A new utility function,
find_free_port, has been introduced intests/kernels/moe/deepep_utils.py. This function dynamically identifies and returns an available network port by binding a temporary socket to port 0. - Improved Test Reliability: The
parallel_launchfunction, used for setting up distributed tests, has been updated to utilize thefind_free_portutility. This ensures thattorch.distributed.init_process_groupalways attempts to use a free port, significantly improving the reliability and robustness of distributed tests by preventing port conflicts.
Using Gemini Code Assist
The full guide for Gemini Code Assist can be found on our documentation page, here are some quick tips.
Invoking Gemini
You can request assistance from Gemini at any point in your pull request via creating an issue comment (i.e. comment on the pull request page) using either /gemini <command> or @gemini-code-assist <command>. Below is a summary of the supported commands.
| Feature | Command | Description |
|---|---|---|
| Code Review | /gemini review |
Performs a code review for the current pull request in its current state. |
| Pull Request Summary | /gemini summary |
Provides a summary of the current pull request in its current state. |
| Comment | @gemini-code-assist | Responds in comments when explicitly tagged, both in issue comments and review comments. |
| Help | /gemini help |
Displays a list of available commands. |
Customization
To customize Gemini Code Assist for GitHub experience, repository maintainers can create a configuration file and/or provide a custom code review style guide (such as PEP-8 for Python) by creating and adding files to a .gemini/ folder in the base of the repository. Detailed instructions can be found here.
Limitations & Feedback
Gemini Code Assist is currently in preview and may make mistakes. Please leave feedback on any instances where its feedback is incorrect or counter productive. You can react with 👍 and 👎 on @gemini-code-assist comments to provide feedback.
You can also get AI-powered code generation, chat, as well as code reviews directly in the IDE at no cost with the Gemini Code Assist IDE Extension.
Footnotes
-
Review the Privacy Notices, Generative AI Prohibited Use Policy, Terms of Service, and learn how to configure Gemini Code Assist in GitHub here. Gemini can make mistakes, so double check it and use code with caution. ↩
There was a problem hiding this comment.
Code Review
The pull request introduces a fix for the address/port already in use error by dynamically finding a free port for the TCP store. It adds a find_free_port function and integrates it into the parallel_launch function. The changes enhance the robustness of the tests by avoiding port conflicts.
Signed-off-by: yewentao256 <zhyanwentao@126.com>
tests/kernels/moe/deepep_utils.py
Outdated
| @@ -79,6 +82,13 @@ def _worker_parallel_launch( | |||
| torch.distributed.destroy_process_group() | |||
|
|
|||
|
|
|||
| def find_free_port(): | |||
There was a problem hiding this comment.
We can probably move this to utils?
There was a problem hiding this comment.
Sounds good, moved to vllm/model_executor/layers/fused_moe/utils.py
Signed-off-by: yewentao256 <zhyanwentao@126.com>
…m-project#20094) Signed-off-by: yewentao256 <zhyanwentao@126.com>
…m-project#20094) Signed-off-by: yewentao256 <zhyanwentao@126.com>
…m-project#20094) Signed-off-by: yewentao256 <zhyanwentao@126.com> Signed-off-by: Will Eaton <weaton@redhat.com>
…m-project#20094) Signed-off-by: yewentao256 <zhyanwentao@126.com>
…m-project#20094) Signed-off-by: yewentao256 <zhyanwentao@126.com>
…m-project#20094) Signed-off-by: yewentao256 <zhyanwentao@126.com> Signed-off-by: avigny <47987522+avigny@users.noreply.github.com>
…m-project#20094) Signed-off-by: yewentao256 <zhyanwentao@126.com>
Purpose
Fix address/port already in use error
Test