[GH-ISSUE #824] [Bug]: Notify should be sent until all tasks of a batch_id have finished. #248

New issue

Closed

opened 2026-10-29 21:11:30 +00:00 by bsakaguchi · 0 comments

bsakaguchi commented

2026-10-29 21:11:30 +00:00

Owner

Originally created by @orrorcol on GitHub (Sep 11, 2025).
Original GitHub issue: https://github.com/kvcache-ai/Mooncake/issues/824

Bug Report

In the mooncake plugin of nixl, checkXfer checks every task's status:
https://github.com/ai-dynamo/nixl/blob/main/src/plugins/mooncake/mooncake_backend.cpp#L296

In the check status api, notify will be sent as soon as a sub-task has finished: https://github.com/kvcache-ai/Mooncake/blob/main/mooncake-transfer-engine/include/transfer_engine.h#L166

This results in that notification is sent before all sub-tasks are finished, which is not what we expected. Besides, return code of sendNotifyByID should also be checked because it may fail.

Before submitting...

Ensure you searched for relevant issues and read the [documentation]

Originally created by @orrorcol on GitHub (Sep 11, 2025). Original GitHub issue: https://github.com/kvcache-ai/Mooncake/issues/824 ### Bug Report In the mooncake plugin of nixl, `checkXfer` checks every task's status: https://github.com/ai-dynamo/nixl/blob/main/src/plugins/mooncake/mooncake_backend.cpp#L296 <img width="1270" height="322" alt="Image" src="https://github.com/user-attachments/assets/a97c8312-62da-4140-bf7d-f5604e97d671" /> In the check status api, notify will be sent as soon as a sub-task has finished: https://github.com/kvcache-ai/Mooncake/blob/main/mooncake-transfer-engine/include/transfer_engine.h#L166 <img width="1002" height="558" alt="Image" src="https://github.com/user-attachments/assets/a2e8942c-a20a-43ab-adce-c42fbc53f25d" /> This results in that notification is sent before all sub-tasks are finished, which is not what we expected. Besides, return code of `sendNotifyByID` should also be checked because it may fail. ### Before submitting... - [ ] Ensure you searched for relevant issues and read the [documentation]

bsakaguchi

2026-10-29 21:11:30 +00:00