fix: use persistent event loop in watcher to prevent loop-closed crash#76
Open
svidskiy wants to merge 2 commits intozilliztech:mainfrom
Open
fix: use persistent event loop in watcher to prevent loop-closed crash#76svidskiy wants to merge 2 commits intozilliztech:mainfrom
svidskiy wants to merge 2 commits intozilliztech:mainfrom
Conversation
Replace asyncio.run() with a dedicated loop.run_until_complete() in the file watcher callback. asyncio.run() creates and closes a new event loop on every call, which breaks async HTTP clients (httpx/ollama/openai/voyage) that cache connections tied to the first loop. Fixes zilliztech#75
This file contains hidden or 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
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
1 participant
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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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.
Summary
Replaces
asyncio.run()with a dedicated persistent event loop (loop.run_until_complete()) in the file watcher callback incore.pyasyncio.run()creates and closes a new event loop on every invocation. Async HTTP clients (httpx, used internally by ollama, openai, voyage) cache connections tied to the first loop. Whenasyncio.run()closes that loop and creates a new one, the cached client attempts to use the dead loop — resulting inRuntimeError: Event loop is closedon the second file change.This is a known httpx limitation — sharing an
AsyncClientacross multipleasyncio.run()calls is not supported. See also httpx#914.Affects all async embedding providers (OpenAI, Ollama, Voyage, Google), not just Ollama as reported. The local provider is unaffected (no HTTP client).
Fixes #75
Test plan
memsearch watch .with Ollama provider, modify a file twice — verify both re-indexes succeed withoutRuntimeErrormemsearch watch .with OpenAI provider, modify a file multiple times — verify no crashuv run python -m pytest) — all 26 tests pass