Skip to content

[PR] ci: register tests/storage-durability.test.ts in LIVE_FILES (#4333) - #4334

Closed
ranxianglei wants to merge 2 commits into
masterfrom
2026-10-22_live-files-storage-durability
Closed

ranxianglei wants to merge 2 commits into
masterfrom
2026-10-22_live-files-storage-durability

Conversation

@ranxianglei

Copy link
Copy Markdown
Owner

(ai: zhipuai-lb/glm-5.3)

#4177 落了 storage durability trio,但新测试 tests/storage-durability.test.ts 用了 child_process.spawn 却没登记 scripts/live-files.mjs 的 LIVE_FILES——#1248 绊线让 master 自身 CI(aa198fe61)与所有基于新 master 的分支全部红。

本 PR 一行修复:按字母序登记该测试(slot 31 / 端口带 19811)。

并发假设已核对(登记前置要求):

  • 仅 spawn node -e "setTimeout() => {}, 60000)" 探针子进程(锁文件第二进程检测),不绑任何端口;
  • 数据目录全部 mkdtempSync 隔离,删除 BILI_STORAGE_UNIFIED 后走自身环境;
  • 本地单跑 5/5 绿。

Fixes #4333

#4177 landed the storage durability trio but its new suite spawns
child processes without registering in scripts/live-files.mjs, so the
#1248 SPAWN_RE tripwire fails every CI run — master aa198fe is red
and every branch rebased onto it inherits the failure.

Concurrency review: the suite only spawns `node -e setTimeout` probes
(no port binding) and isolates data in mkdtempSync dirs; slot 31 /
zone band 19811 is free. Fixes #4333.
@ranxianglei ranxianglei changed the title ci: register tests/storage-durability.test.ts in LIVE_FILES (#4333) [PR] ci: register tests/storage-durability.test.ts in LIVE_FILES (#4333) Oct 22, 2026
@github-actions

github-actions Bot commented Oct 22, 2026 •

Copy link
Copy Markdown
Contributor

📦 Built Package Artifact

Branch: 2026-10-22_live-files-storage-durability (0f21448)

Option A — Install this exact build (recommended)

npm install -g billion-context@0.0.5142-pr.4334.2f0f160

This exact version is immutable and reproducible. The rolling pr tag tracks the newest test build across all PRs.

Option B — Download artifact

  1. Download the artifact from the Actions run
  2. Extract the tarball and install:
tar xzf billion-context-pr4334.tgz
npm install -g package

Install instructions are refreshed on each push.

Since #4177 the storage contract suites transitively import acp-kernel
at runtime (backend -> store -> runtime-status -> knobs -> config), and
tsx resolves it to kernel/dist — which this job never built. Mirror the
prerequisite the main test matrix (run-ci-tests pretest parity) and
ci-e2e-dsh.yml already carry. Fixes #4336.
@ranxianglei

Copy link
Copy Markdown
Owner Author

(ai: zhipuai-lb/glm-5.3)

追加第二处 master 红:mysql-contract 车道(#4336)——#4177 给 store.ts 新加的 runtime-status import 把 acp-kernel 运行时依赖带进了 storage 契约测试链(backend → store → runtime-status → knobs → config → acp-kernel),而 mysql-contract job 从不构建 kernel/dist,1da8eca6 起红(c1d7eab 前全绿)。

修法=照抄 ci-e2e-dsh.yml:90 的已验证模式,job 里补一行 npm --prefix kernel run build(与主测试矩阵 run-ci-tests.mjs 的 pretest parity 同语义)。

本 PR 现在两件事:①LIVE_FILES 登记(Fixes #4333)②mysql-contract kernel 构建(Fixes #4336)。两个都是从 #4177 继承的 master CI 红,合并即恢复 master 全绿。

@ranxianglei

Copy link
Copy Markdown
Owner Author

(ai: zhipuai-lb/glm-5.3)
更新:本 PR 的两个 commit 都有了归宿——

本 PR 可关闭(superseded by #4350)。另:#4334 头 job 上 windows-22-1 的两条 EBUSY 失败与本 PR 改动无关——master 9e2d217(不含本 PR 任何改动)同签名失败,根因与修复见 #4347。

@ranxianglei

Copy link
Copy Markdown
Owner Author

🤖 Powered by ework · qwen3.8-27b

(ai: vllm-qwen/qwen3.8-27b)

判定:不在本分支修 —— 本 PR 已被完全取代(两个 commit 均冗余),红车道属于在途跟踪的存量缺陷。按自动修复车道规则(每 PR 最多 2 次;失败需要设计决策或不是小修时发判定并停),本轮到判定为止。

1. 是什么问题
test (windows-latest, 22, 1)(head 0f21448,job 119964230446)挂了两条 corrupt-index.db 降级测试:tests/storage-engine-proofs.test.ts:381(#2771)和 tests/web-storage-banner.test.ts:302(#3040)。断言全部通过,死在清理阶段 rmrf 删临时目录:EBUSY: resource busy or locked, unlink '...\Temp\bili-*\index.db',重试梯子(~9.6s)烧尽后抛出。

2. 什么原因导致的
不是本 PR 的改动引起——逐项验证过:

机制有两个竞争假设,都已立项跟踪:

3. 影响范围与严重程度
纯 CI/测试层问题,最终用户无运行时影响。但按 #4338 的确定性判断,windows shard 1 恒红期间所有 PR 合并被阻塞(rerun 无效)。

4. 怎么处理的(为什么不直接修)

5. 是否有回归问题
未做任何代码改动,无回归面。对失败 lane 的既有行为审计:两条测试的断言本体在所有观测运行中均通过(降级行为正确),红的只有清理阶段。

6. 是否破坏缓存
不涉及。

7. 是否改了设计方向
没有。留一个待拍板项供参考:若 #4338「确定性句柄泄漏」成立,#4350 的有界重试会再烧一个分片复现同红(不掩盖真回归,但也治不了病),根因线(最小 Windows repro → driver 层显式释放,见 #4338 修法方向方向①–④)仍需推进;若 #4347「AV 瞬态」成立,#4350 即闭环。两 issue 是同一症状同一族,建议合并到一处裁决。

8. 是否引入新配置
无。

判定:#4334 关闭(superseded by master 133cdd2 + PR #4350);windows-22-shard-1 红车道等 #4350 的 CI 结果 + owner 对 #4338/#4347 机制之争的裁决。

@ranxianglei ranxianglei added the duplicate This issue or pull request already exists label Oct 23, 2026
yanhuojiyusheng pushed a commit to yanhuojiyusheng/billion-context that referenced this pull request Oct 23, 2026
)

Since ranxianglei#4177 the storage suites transitively import acp-kernel at runtime
(store.ts -> runtime-status -> knobs -> config); tsx resolves it to the
built dist, so the mysql-contract job dies with ERR_MODULE_NOT_FOUND on a
fresh runner. Same prerequisite the main test matrix already has via
run-ci-tests' pretest step. Cherry-picked from ranxianglei#4334.
yanhuojiyusheng pushed a commit to yanhuojiyusheng/billion-context that referenced this pull request Oct 23, 2026
…glei#4347)

Fixes ranxianglei#4347. The windows-22-1 lane hit the ranxianglei#2799/ranxianglei#2908 EBUSY family twice
in 24h (PR ranxianglei#4334 head job + master 9e2d217, byte-identical signatures):
Windows AV/indexer briefly locks a freshly-written index.db past
tests/tmp-rm.ts' ~9.6s retry ladder, failing the two corrupt-index.db
tests with 'EBUSY: resource busy or locked, unlink ...'. The classifier
judged that concrete -> red, no retry, blocking unrelated PRs.

Adds a third recoverable shape 'cleanup-lock' beside ranxianglei#3029 (swallowed
deaths) and ranxianglei#3665 (event-loop drain): when EVERY failing subtest carries
a transient-lock errno on an unlink/rmdir through a rimraf teardown
stack, retry the whole shard once and exit 0 only if the retry is green.
A deterministic lock (a genuine handle leak) reproduces on the retry and
still fails, so no regression can be masked; a raw user-code unlinkSync
EBUSY (no rimraf frames) stays concrete and unretried. Both real failed
logs classify cleanup-lock; the real ranxianglei#4336 regression log stays
concrete. Unit coverage asserts both directions (ranxianglei#3617).
ranxianglei added a commit that referenced this pull request Oct 24, 2026
…_FILES (#4449 follow-up)

#4449 landed two process-spawning suites without registering them, so
every test shard now aborts at startup ('process-spawning test files
missing from LIVE_FILES') on master and all open PRs. Same disease as
#4333/#4334 — register them.
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

duplicate This issue or pull request already exists

Projects

None yet

1 participant