Revert "[PD] Make pending reqs resolving more robust"#20779
Conversation
This reverts commit 738cbde.
Summary of ChangesHello, 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 reverts the changes introduced in a previous PR (#20505) that aimed to enhance the robustness of pending request resolution. The decision to revert was made because the underlying root cause of the issue has been addressed in a separate, more comprehensive bugfix (#20756). Consequently, the retry mechanisms and associated state management for fetching prefill parallel information and resolving pending requests have been removed. Highlights
🧠 New Feature in Public Preview: You can now enable Memory to help Gemini Code Assist learn from your team's feedback. This makes future code reviews more consistent and personalized to your project's style. Click here to enable Memory in your admin console. Changelog
Using Gemini Code AssistThe 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 by creating a comment using either
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 Limitations & Feedback Gemini Code Assist 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. If you're interested in giving your feedback about your experience with Gemini Code Assist for GitHub and other Google products, sign up here. 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
|
|
/rerun-stage stage-c-test-8-gpu-h20 |
|
✅ Triggered |
![gemini-code-assist[bot]](https://avatars.githubusercontent.com/in/956858?s=60&v=4){kind=small}
There was a problem hiding this comment.
Code Review
This pull request appears to be a revert of a previous change, effectively refactoring the retry logic for fetching prefill server information. The logic is moved from decode.py back into a centralized ensure_parallel_info method in conn.py. This is a good change as it centralizes the logic and simplifies the calling code. The changes also fix a bug where record_failure was called with an incorrect number of arguments. The overall changes look good and improve code quality.
| def ensure_parallel_info( | ||
| self, bootstrap_addr: str, max_retries: int = 5, retry_interval: float = 1.0 | ||
| ) -> bool: |
There was a problem hiding this comment.
The retry parameters max_retries and retry_interval are hardcoded as default values. It would be more flexible to make them configurable through environment variables, similar to other parameters in this file (e.g., SGLANG_DISAGGREGATION_BOOTSTRAP_TIMEOUT). This would allow for easier tuning in different deployment environments without changing the code.
CI has passed. |
Revert "[PD] Make pending reqs resolving more robust" (sgl-project#20779)
1 participant
Reverts #20505
Because the root cause and bugfix are detailed here #20756