[BugFix] Overhaul async request cancellation #7111
Merged
Conversation
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
|
👋 Hi! Thank you for contributing to the vLLM project. Once the PR is approved and ready to go, please make sure to run full CI as it is required to merge (or just use auto-merge). To run full CI, you can do one of these:
🚀 |
|
Nice |
…borts # Conflicts: # vllm/entrypoints/api_server.py
DarkLight1337
approved these changes
Aug 7, 2024
njhill
added a commit
to njhill/vllm
that referenced
this pull request
Aug 7, 2024
vllm-project#7111 made a change to the merge_async_iterators utils function to add an is_cancelled arg. It would be good for this new arg to be optional to retain backwards compatibility for other server front-ends that might already be using this utility function.
njhill
added a commit
to njhill/vllm
that referenced
this pull request
Aug 7, 2024
vllm-project#7111 made a change to the merge_async_iterators utils function to add an is_cancelled arg. It would be good for this new arg to be optional to retain backwards compatibility for other server front-ends that might already be using this utility function.
njhill
added a commit
to njhill/vllm
that referenced
this pull request
Aug 9, 2024
Follow-on from vllm-project#7111, avoid unecessary enqueuing a final message after an exception and avoid aborting requests in the engine that were never started.
sfc-gh-mkeralapura
pushed a commit
to sfc-gh-mkeralapura/vllm
that referenced
this pull request
Aug 12, 2024
kylesayrs
pushed a commit
to neuralmagic/vllm
that referenced
this pull request
Aug 17, 2024
fialhocoelho
pushed a commit
to opendatahub-io/vllm
that referenced
this pull request
Aug 22, 2024
1 task
Alvant
pushed a commit
to compressa-ai/vllm
that referenced
this pull request
Oct 26, 2024
Signed-off-by: Alvant <alvasian@yandex.ru>
KuntaiDu
pushed a commit
to KuntaiDu/vllm
that referenced
this pull request
Nov 20, 2024
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
There are a number of problems currently with how request cancellation works upon client disconnection in the openai api server front-end and
AsyncLLMEngine:This is a problem for production resilience and has a compounding affect when the server is overloaded and client requests time-out, with the server continuing to do useless work.
This PR reworks how the cancellation is propagated to make it more robust and consistent:
asyncio.aclose()) rather than async iterators in most casesmerge_async_iteratorsto encapsulate polling for disconnection, even before any results have been produced (while the request is queued)iterate_with_cancellationfunction used for the single-prompt request casesAsyncLLMEnginedifferentiate between cancelled requests and those that finish normally (generator returned from thegenerate()will now raise aCancelledErrorin the former case)I also plan to add some new tests to cover these various cases.