[Cherry-Pick][BugFix] Fix bugs in /v1/abort_requests interface from PR(#6992) (#7176)

[qwes5s5]

Motivation

💡 If this PR is a Cherry Pick, the PR title needs to follow the format by adding the [Cherry-Pick] label at the very beginning and appending the original PR ID at the end. For example, [Cherry-Pick][CI] Add check trigger and logic(#5191)

💡 如若此PR是Cherry Pick，PR标题需遵循格式，在最开始加上[Cherry-Pick]标签，以及最后面加上原PR ID，例如[Cherry-Pick][CI] Add check trigger and logic(#5191)

Modifications

Usage or Command

Accuracy Tests

Checklist

• Add at least a tag in the PR title.
  • Tag list: [[FDConfig],[APIServer],[Engine], [Scheduler], [PD Disaggregation], [Executor], [Graph Optimization], [Speculative Decoding], [RL], [Models], [Quantization], [Loader], [OP], [KVCache], [DataProcessor], [BugFix], [Docs], [CI], [Optimization], [Feature], [Benchmark], [Others], [XPU], [HPU], [GCU], [DCU], [Iluvatar], [Metax]]
  • You can add new tags based on the PR content, but the semantics must be clear.
• Format your code, run pre-commit before commit.
• Add unit tests. Please write the reason in this PR if no unit tests.
• Provide accuracy results.
• If the current PR is submitting to the release branch, make sure the PR has been submitted to the develop branch, then cherry-pick it to the release branch with the [Cherry-Pick] PR tag.

[paddle-bot]

Thanks for your contribution!

[PaddlePaddle-bot]

🤖 AI Code Review | 2026-04-22 11:37:08

📋 Review 摘要

PR 概述：修复 /v1/abort_requests 接口中的多个 Bug，包括 EOS token 访问方式、abort 等待逻辑中的竞态处理以及 finish_reason 枚举值缺失问题。
变更范围：engine/common_engine.py、engine/sched/resource_manager_v1.py、entrypoints/openai/protocol.py、tests/engine/test_common_engine.py
影响面 Tag：[Engine] [APIServer] [Scheduler]

📝 PR 规范检查

PR 描述中 Motivation / Modifications 均未填写，无法从 PR body 了解具体 Bug 背景和变更影响面，请补充以下内容：

Motivation 建议（可直接复制）：

## Motivation
修复 `/v1/abort_requests` 接口中的多个 Bug：
1. `_control_abort_requests` 通过 `self.data_processor.eos_token_ids[0]` 获取 EOS token 在某些情况下会抛出 AttributeError；
2. `_wait_abort_complete` 在请求已自然结束时仍等待 abort 完成，导致不必要的超时；
3. `recycle_abort_task` / `_trigger_abort` 中使用 `set.remove` 在 key 不存在时抛出 KeyError；
4. `finish_requests` 未清理 abort 相关集合，可能导致后续误判；
5. OpenAI 协议中 `finish_reason` 缺少 `"abort"` 枚举值导致 Pydantic 校验报错。

## Modifications
- 使用 `getattr(request, "eos_token_ids", [0])` 安全获取 EOS token；
- `_wait_abort_complete` 入口及循环内增加"已完成请求"快速跳过逻辑；
- 将 `set.remove` 替换为 `set.discard` 避免 KeyError；
- `finish_requests` 中同步清理 abort 相关集合；
- 四个 Response/Choice 模型中补充 `"abort"` 枚举值；
- 更新相关单元测试以匹配新逻辑。

---

问题

级别	文件	概述
🟡 建议	common_engine.py:1590	getattr 默认值 [0] 使用魔法数字，建议添加注释或 warning 日志说明 fallback 原因
❓ 疑问	common_engine.py:1641	循环内无锁读取 alive_reqs 并直接操作 abort 集合，线程安全性待确认

---

总体评价

本 PR 修复了 abort 接口中多处潜在的 KeyError 和属性访问异常，逻辑方向正确，测试也同步更新。核心逻辑中的无锁并发访问点建议作者确认线程安全性，同时 PR 描述需补充 Motivation/Modifications 以满足项目规范。

[codecov-commenter]

Codecov Report

❌ Patch coverage is 66.66667% with 6 lines in your changes missing coverage. Please review.
⚠️ Please upload report for BASE (release/2.6@b0fde16). Learn more about missing BASE report.

Files with missing lines	Patch %	Lines
fastdeploy/engine/common_engine.py	40.00%	5 Missing and 1 partial ⚠️

Additional details and impacted files

@@              Coverage Diff               @@
##             release/2.6    #7551   +/-   ##
==============================================
  Coverage               ?   73.75%           
==============================================
  Files                  ?      376           
  Lines                  ?    53082           
  Branches               ?     8298           
==============================================
  Hits                   ?    39152           
  Misses                 ?    11184           
  Partials               ?     2746           

Flag	Coverage Δ
GPU	73.75% <66.66%> (?)

Flags with carried forward coverage won't be shown. Click here to find out more.

☔ View full report in Codecov by Sentry.
📢 Have feedback on the report? Share it here.

🚀 New features to boost your workflow:

• ❄️ Test Analytics: Detect flaky tests, report on failures, and find test suite problems.

[qwes5s5]

/re-run stable_tests