fix: harden storage semantics #4118
Merged
+516
−211
Conversation
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
Fixes issues in the storage system by guaranteeing immediate durability for responses and ensuring background writers stay alive. Responses to the OpenAI-compatible API now write directly to Postgres/SQLite inside the request instead of detouring through an async queue that might never drain; this restores the expected read-after-write behavior and removes the "response not found" races reported by users. The access-control shim was stamping owner_principal/access_attributes as SQL NULL, which Postgres interprets as non-public rows; fixing it to use the empty-string/JSON-null pattern means conversations and responses stored without an authenticated user stay queryable (matching SQLite). The inference-store queue remains for batching, but its worker tasks now start lazily on the live event loop so server startup doesn't cancel them—writes keep flowing even when the stack is launched via llama stack run.
ashwinb
requested review from
bbrowning,
ehhuang,
franciscojavierarceo,
hardikjshah,
leseb,
mattf,
raghotham,
reluctantfuturist,
slekkala1,
terrytangyuan and
yanxi0830
as code owners
meta-cla
bot
added
the
CLA Signed
ashwinb
commented
Nov 10, 2025
| responses_store=postgres_config, | ||
| ), | ||
| config=MetaReferenceAgentsImplConfig( | ||
| persistence=AgentPersistenceConfig( |
Contributor
Author
There was a problem hiding this comment.
important bug fix, unrelated to this one but really the postgres definition was all kinds of wrong
ehhuang
approved these changes
Nov 11, 2025
leseb
approved these changes
Nov 12, 2025
Contributor
Author
|
@Mergifyio backport release-0.3.x |
✅ Backports have been created
|
mergify bot
pushed a commit
that referenced
this pull request
Nov 12, 2025
Fixes issues in the storage system by guaranteeing immediate durability for responses and ensuring background writers stay alive. Three related fixes: * Responses to the OpenAI-compatible API now write directly to Postgres/SQLite inside the request instead of detouring through an async queue that might never drain; this restores the expected read-after-write behavior and removes the "response not found" races reported by users. * The access-control shim was stamping owner_principal/access_attributes as SQL NULL, which Postgres interprets as non-public rows; fixing it to use the empty-string/JSON-null pattern means conversations and responses stored without an authenticated user stay queryable (matching SQLite). * The inference-store queue remains for batching, but its worker tasks now start lazily on the live event loop so server startup doesn't cancel them—writes keep flowing even when the stack is launched via llama stack run. Closes #4115 ### Test Plan Added a matrix entry to test our "base" suite against Postgres as the store. (cherry picked from commit 492f79c) # Conflicts: # .github/workflows/integration-tests.yml # llama_stack/distributions/ci-tests/run-with-postgres-store.yaml # llama_stack/distributions/starter-gpu/run.yaml # llama_stack/distributions/starter/run.yaml # llama_stack/distributions/starter/starter.py # llama_stack/providers/utils/inference/inference_store.py # llama_stack/providers/utils/responses/responses_store.py # tests/integration/ci_matrix.json
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.
Fixes issues in the storage system by guaranteeing immediate durability for responses and ensuring background writers stay alive. Three related fixes:
Responses to the OpenAI-compatible API now write directly to Postgres/SQLite inside the request instead of detouring through an async queue that might never drain; this restores the expected read-after-write behavior and removes the "response not found" races reported by users.
The access-control shim was stamping owner_principal/access_attributes as SQL NULL, which Postgres interprets as non-public rows; fixing it to use the empty-string/JSON-null pattern means conversations and responses stored without an authenticated user stay queryable (matching SQLite).
The inference-store queue remains for batching, but its worker tasks now start lazily on the live event loop so server startup doesn't cancel them—writes keep flowing even when the stack is launched via llama stack run.
Closes #4115
Test Plan
Added a matrix entry to test our "base" suite against Postgres as the store.