[PD] Remove unused server args for disaggregation

[ShangmingCai]

Motivation

We don't need these server args (since decode nodes can get these info from bootstrap servers, prefill can get these info from register info), they are not used for all backends, and will cause misunderstanding for open-source users.

Modifications

Accuracy Tests

Benchmarking and Profiling

Checklist

• Format your code according to the Format code with pre-commit.
• Add unit tests according to the Run and add unit tests.
• Update documentation according to Write documentations.
• Provide accuracy and speed benchmark results according to Test the accuracy and Benchmark the speed.
• Follow the SGLang code style guidance.

Review Process

1. Ping Merge Oncalls to start the PR flow. See the PR Merge Process.
2. Get approvals from CODEOWNERS and other reviewers.
3. Trigger CI tests with comments or contact authorized users to do so.
   • /tag-run-ci-label, /rerun-failed-ci, /tag-and-rerun-ci
4. After green CI and required approvals, ask Merge Oncalls to merge.

[gemini-code-assist]

Summary of Changes

Hello, 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 streamlines the disaggregation feature by removing obsolete server configuration parameters and their corresponding code references. The changes simplify the codebase and documentation, making the disaggregation setup cleaner and easier to manage without impacting functionality, as the removed arguments were no longer in use.

Highlights

• Removed Unused Server Arguments: The pull request eliminates several server arguments related to disaggregation, specifically --disaggregation-decode-tp, --disaggregation-decode-dp, and --disaggregation-prefill-pp, which were no longer utilized in the system.
• Codebase Cleanup: Associated code references to these removed arguments have been cleaned up across various files, including the KVArgs class, DecodeServer and PrefillServer initializations, and the Scheduler's disaggregation setup.
• Documentation Update: The documentation for server arguments has been updated to reflect the removal of the obsolete disaggregation parameters, ensuring accuracy and reducing confusion for users.

🧠 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

• docs/advanced_features/server_arguments.md
  • Removed documentation entries for --disaggregation-decode-tp, --disaggregation-decode-dp, and --disaggregation-prefill-pp.
• python/sglang/srt/disaggregation/base/conn.py
  • Removed the decode_tp_size attribute from the KVArgs class.
• python/sglang/srt/disaggregation/decode.py
  • Removed prefill_pp_size from the __init__ method parameters and its subsequent assignments.
  • Removed the assignment of kv_args.prefill_pp_size in _init_kv_manager.
• python/sglang/srt/disaggregation/prefill.py
  • Removed decode_tp_size and decode_dp_size from the __init__ method parameters and their subsequent assignments.
  • Removed the assignments of kv_args.decode_tp_size and kv_args.prefill_pp_size in _init_kv_manager.
• python/sglang/srt/managers/scheduler.py
  • Removed the prefill_pp_size argument during DecodeServer initialization.
  • Removed the decode_tp_size and decode_dp_size arguments during PrefillServer initialization.
• python/sglang/srt/server_args.py
  • Removed disaggregation_decode_tp, disaggregation_decode_dp, and disaggregation_prefill_pp attributes from the ServerArgs class.
  • Removed assertions and assignments related to these arguments within the _handle_pd_disaggregation method.
  • Removed the add_argument calls for these parameters in the add_cli_args function.
  • Removed the validate_disagg_tp_size method.

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 by creating a comment using either /gemini <command> or @gemini-code-assist <command>. Below is a summary of the supported commands on the current page.

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 pull request 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 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

1. 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. ↩

[gemini-code-assist]

Code Review

This pull request effectively removes several unused server arguments related to disaggregation, which simplifies the configuration. The changes are consistently applied across the argument parsing logic, documentation, and their usage sites in the scheduler and disaggregation modules. I have one minor suggestion to ensure the removal is complete and to prevent potential issues.

I am having trouble creating individual review comments. Click here to see my feedback.

python/sglang/srt/disaggregation/prefill.py (138-139)

While you've removed the assignments to kv_args.decode_tp_size and kv_args.prefill_pp_size, it seems prefill_pp_size might still be present in the KVArgs class definition in python/sglang/srt/disaggregation/base/conn.py. For completeness and to avoid potential AttributeError if it's accessed elsewhere unexpectedly, consider removing prefill_pp_size from the KVArgs class definition as well.

[ShangmingCai]

/rerun-stage stage-c-test-8-gpu-h20

[github-actions]

✅ Triggered stage-c-test-8-gpu-h20 to run independently (skipping dependencies).

[github-actions]

🔗 View workflow run

[ShangmingCai]

/tag-and-rerun-ci

[ShangmingCai]

Related CI has passed.

[iyastreb]

@ShangmingCai This PR breaks NIXL plugin, where decode_tp_size is still referenced
It seems Mori is also affected
Could you please fix it?