[WORKFLOW SDK FEATURE REQUEST] Retry WaitForInstanceCompletion/Start on a server-sent CANCELLED

未关闭

#1,246 创建于 2026年9月28日 0 条评论 0 个 reaction 已指派 0 人在 GitHub 查看

维护者通常 3 天内回复

还没有人认领这个 Issue。

评估

难度: 4/5
预计耗时: 3-5 天
新手友好度: 55/100

Issue 类型: 功能
描述清晰度: 基本清楚
活跃度: 活跃
技术栈: python
领域: backend, distributed-systems

调研方向

Start in dapr/ext/workflow/_durabletask/client.py around lines 335-349 and the matching retry logic in aio/client.py. Read the three cancellation tests in tests/ext/workflow/durabletask/test_orchestration_wait.py and test_client_async.py, then confirm the intended retry and timeout behavior with maintainers. Done means both clients re-issue eligible waits after server CANCELLED responses while preserving caller deadlines and the updated tests pass.

由索引模型根据 Issue 内容生成。

描述

dapr-ext-workflow kind/enhancement

Proposal

wait_for_workflow_completion and wait_for_workflow_start (sync and aio) should re-issue WaitForInstanceCompletion / WaitForInstanceStart when the server ends the call with CANCELLED and the caller's own timeout hasn't expired. Today that status reaches the caller as an error, even though the workflow is still running. This asks maintainers to decide on the behaviour first, because it reverses assertions added in #1112. It's a proposal, not a PR.

Why the server sends CANCELLED

When daprd shuts down or restarts mid-wait (rollout, pod restart, hot-reload, a config change that restarts the runtime), the actors router cancels every in-flight call it tracks. The wait then returns CANCELLED "context canceled", even though the caller is still waiting. The mechanism and a repro are in dapr/dapr#10566.

Reproduced against native daprd 1.18.0:

a 60 s workflow with wait_for_workflow_completion(id) and no timeout returns normally;
the same run with daprd sent SIGTERM about 10 s in fails with StatusCode.CANCELLED "context canceled".

Why retrying is safe

A blocking unary call only gets CANCELLED from the server. The sync client can't cancel it mid-flight, and in the aio client a caller's cancellation raises asyncio.CancelledError, not AioRpcError. A client-side deadline shows up as DEADLINE_EXCEEDED, which already maps to TimeoutError.
The wait only reads state, so re-issuing it is idempotent. It returns straight away if the workflow already finished.

What would change

Add CANCELLED to the retried codes, when the caller's timeout hasn't expired, in _durabletask/client.py L335 and _durabletask/aio/client.py L230. _is_deadline_cancellation (L349) already turns a CANCELLED after the deadline into TimeoutError.
Restart the 30 s transient-retry window after each successful re-issue, rather than anchoring it on the first transient error. Otherwise a sidecar that restarts more than once during a long wait would still fail it.
These tests from #1112 currently assert that a CANCELLED reaches the caller, and would flip:

Relationship to the runtime fix

The proper fix is in daprd: return UNAVAILABLE in this case (dapr/dapr#10566). The SDK already retries UNAVAILABLE. This change would protect users on current and older runtimes until that ships, and it would still help afterwards with proxies that reset the stream.

Question for maintainers

Is it acceptable to treat a server-sent CANCELLED on these two waits as retryable, and change the #1112 tests? If yes, it's a small change in both clients plus tests.

主要语言: Python
星标: 272
派生: 152
平均合并: 3 天 13 小时
30 天内合并 PR: 11

环境准备

阅读贡献指南

从这里开始

先读完整个 Issue，再读项目的贡献指南。
在 Issue 下留言说明你要接手 —— 这能避免两个人做同样的事。
Fork 仓库，在一个分支上完成修改。
提交 Pull Request，并在描述里引用这个 Issue 编号。

dapr/python-sdk 的其他 Issue

SDK Alignment of Workflow Examples
未关闭
- 哪里不对？
dapr-ext-workflow good first issue kind/enhancement P2

难度 2/5 1-2 天新手友好度 72/100

dapr/python-sdk#853 · 4 条评论 · 2025年10月21日创建

维护者通常 3 天内回复
[BUG] subscribe_with_handler stops silently: async task can die or be collected, sync retry path exits, handler errors reconnect the stream
未关闭
- 哪里不对？
kind/bug

难度 4/5 3-5 天新手友好度 65/100

dapr/python-sdk#1233 · 2026年9月24日创建

维护者通常 3 天内回复
[BUG] Configuration subscription stops permanently after the daprd sidecar restarts
未关闭
- 哪里不对？
kind/bug

难度 4/5 3-5 天新手友好度 68/100

dapr/python-sdk#1232 · 2026年9月24日创建

维护者通常 3 天内回复
[BUG] Streaming subscribe returns None after a reconnect, making test_streaming_subscribe_receives_published_message flaky
未关闭
- 哪里不对？
kind/bug

难度 4/5 3-5 天新手友好度 45/100

dapr/python-sdk#1230 · 2026年9月24日创建

维护者通常 3 天内回复
[WORKFLOW SDK BUG] Workflow constructors reject Unix-domain socket DAPR_GRPC_ENDPOINT
未关闭
- 哪里不对？
dapr-ext-workflow kind/enhancement

难度 3/5 1-2 天新手友好度 78/100

dapr/python-sdk#1213 · 2026年9月14日创建

维护者通常 3 天内回复

查看 dapr/python-sdk 的全部 Issue

相似的 Issue

Reject out-of-range --seed before the run starts
未关闭
- 哪里不对？
难度 2/5 1-3 小时新手友好度 84/100

PedestrianDynamics/pyFDS-Evac#343 · 2026年9月29日创建

维护者通常 1 天内回复
test_run_with_command_flags fails on macOS (BSD printenv doesn't support --version)
未关闭
- 哪里不对？
难度 2/5 1-3 小时新手友好度 88/100

theskumar/python-dotenv#708 · 2026年9月29日创建
gdr_decode_packed_bf16 accepts NaN fixed scale and returns NaN output
未关闭
- 哪里不对？
难度 1/5 1 小时以内新手友好度 88/100

ROCm/aiter#5951 · 2026年9月29日创建

维护者通常 2 天内回复
DOC: document unit-preserving behaviour in timedelta/datetime arithmetic
未关闭
- 哪里不对？
Docs Timedelta

难度 2/5 1-3 小时新手友好度 72/100

pandas-dev/pandas#69919 · 2026年9月29日创建

维护者通常 1 天内回复
Document why west has two ways to access west.configuration.Configuration
未关闭
- 哪里不对？
API documentation

难度 2/5 1-3 小时新手友好度 72/100

zephyrproject-rtos/west#1009 · 2 条评论 · 2026年9月29日创建

维护者通常 3 天内回复

更多 Python Issue

[WORKFLOW SDK FEATURE REQUEST] Retry WaitForInstanceCompletion/Start on a server-sent CANCELLED

哪里不对？

评估

调研方向

描述

Proposal

Why the server sends CANCELLED

Why retrying is safe

What would change

Relationship to the runtime fix

Question for maintainers

环境准备

从这里开始

dapr/python-sdk 的其他 Issue

哪里不对？

哪里不对？

哪里不对？

哪里不对？

哪里不对？

相似的 Issue

哪里不对？

哪里不对？

哪里不对？

哪里不对？

哪里不对？

把新 issue 发到你的邮箱