workflow-runtime 0.0.3: sweep() never recovers a run whose worker died mid-step (fixed on main by #15, unreleased)
還沒有人認領這個 Issue。
評估
- 難度
- 1/5
- 預估耗時
- 1 小時以內
- 新手友好度
- 15/100
- Issue 類型
- 功能
- 描述清晰度
- 描述清楚
- 活躍度
- 活躍
- 技術堆疊
- github, node.js, typescript
研究方向
This asks for release PR #16 to publish core 0.0.5 and runtime 0.0.4; the fix is already merged in #15 and verified on main. Read #16 for the release status and project release instructions, then check whether the requested versions have shipped to npm. Done means both published packages include the fix so users can pin released versions.
由索引模型根據 Issue 內容生成。
描述
Versions
@tanstack/workflow-core 0.0.4, @tanstack/workflow-runtime 0.0.3 (latest on npm)
Describe the bug and the steps to reproduce it
If the process driving a run dies inside a step, the run stays running with an expired lease forever. In 0.0.3, runtime.sweep() only handles schedules and timers and never calls store.claimStaleRuns. Calling claimStaleRuns plus runWorkflow by hand doesn't help either, because core 0.0.4 has no recover mode, so replay stops at the last checkpoint.
#15 (merged 2026-07-21) fixes this: expired-run recovery in sweep(), recover: true in core, and lease heartbeats. I built main at b928717 and confirmed sweep() re-runs the interrupted step and finishes the run, and that the heartbeat stops a long step from being reclaimed mid-flight.
Request: could the release PR #16 be merged, so core 0.0.5 and runtime 0.0.4 ship to npm? That would let us pin a published version instead of a git SHA.
Your Minimal, Reproducible Example
import { createWorkflow } from '@tanstack/workflow-core'
import { defineWorkflowRuntime, inMemoryWorkflowExecutionStore } from '@tanstack/workflow-runtime'
const wf = createWorkflow({ id: 'wf', version: 'v1' }).handler(async (ctx) => {
await ctx.step('a', async () => 1)
return 'done'
})
const store = inMemoryWorkflowExecutionStore()
const runtime = defineWorkflowRuntime({ store, workflows: { wf: { load: async () => wf } }, defaultLeaseMs: 50 })
await store.createRun({ runId: 'r1', workflowId: 'wf', workflowVersion: 'v1', input: {}, now: Date.now() })
await store.claimRun({ runId: 'r1', leaseOwner: 'dead-worker', leaseMs: 50, now: Date.now() }) // worker dies here
await new Promise((r) => setTimeout(r, 100))
const sweep = await runtime.sweep()
console.log(sweep.summary, (await store.loadRun('r1'))?.status)
// 0.0.3: { scheduled: {}, timers: {}, eventCount: 0 } 'running' (expected: recovered, then finished)
Do you intend to try to help solve this bug with your own PR?
No — the fix is already on main; this is a request to publish it.
Terms & Code of Conduct
- I agree to follow this project's Code of Conduct
- I understand that if my bug cannot be reliably reproduced in a debuggable environment, it will probably not be fixed and this issue may even be closed.
- 主要語言
- TypeScript
- 星號
- 221
- 分支
- 10
- PR 合併指標
- 30 天內沒有已合併 PR
環境準備
- 沒有 Dockerfile 或 Docker Compose 檔案
- 有 Pull Request 範本
- 閱讀貢獻指南
從這裡開始
- 先讀完整個 Issue，再讀專案的貢獻指南。
- 在 Issue 下留言說明你要接手 —— 這能避免兩個人做同樣的事。
- Fork 儲存庫，在一個分支上完成修改。
- 送出 Pull Request，並在描述裡引用這個 Issue 編號。
TanStack/workflow 的其他 Issue
-
難度 4/5 1-2 天 新手友好度 52/100
-
難度 5/5 一週以上 新手友好度 25/100
-
難度 5/5 一週以上 新手友好度 20/100
查看 TanStack/workflow 的全部 Issue
相似的 Issue
-
enhancement
難度 2/5 1-3 小時 新手友好度 73/100
ajv-validator/ajv#2689 ·
-
[Bug]: IME confirmation Enter can submit the SSH port-forward dialog可能已有人在做 @seijikohara 今天認領。 未關閉
難度 2/5 1-3 小時 新手友好度 88/100
維護者通常 1 天內回覆
-
難度 2/5 1-3 小時 新手友好度 72/100
home-assistant/frontend#55323 ·
維護者通常 1 天內回覆
-
community first-timers-only good first issue hacktoberfest help wanted low hanging fruit up-for-grabs
難度 1/5 1 小時以內 新手友好度 72/100
lingdojo/kana-dojo#33188 · 1 則留言 · 5 個 reaction ·
維護者通常 1 天內回覆
-
self-driving
難度 1/5 1-3 小時 新手友好度 86/100
維護者通常 1 天內回覆