Skip to content

fix(kernel): honor the client CloudFetch default - #573

Closed
aakash-saravanan-db wants to merge 1 commit into
databricks:mainfrom
aakash-saravanan-db:aakash-saravanan_data/fix/cloudfetch-disabled
Closed

aakash-saravanan-db wants to merge 1 commit into
databricks:mainfrom
aakash-saravanan-db:aakash-saravanan_data/fix/cloudfetch-disabled

Conversation

@aakash-saravanan-db

Copy link
Copy Markdown
Collaborator

The kernel backend ignores the DBSQLClient useCloudFetch default, so applications still download results from cloud storage when they disable it. Forward useCloudFetch: false when opening a session, preserving caller configuration and giving the disable option precedence over differently cased capability keys. Per-statement overrides remain unsupported on the kernel backend.

Validation: build, lint, type check, Prettier, and 1,510 unit tests passed. A real driver loaded the locally built dist and fetched all 250,000 rows with 32,000,000 payload bytes and the expected data hash. A proxy observed zero storage downloads when disabled and two when enabled. Matched Thrift comparisons, error recovery, reuse, and close passed for connection configuration.

Signed-off-by: Aakash Saravanan <aakash.saravanan@databricks.com>
@aakash-saravanan-db

Copy link
Copy Markdown
Collaborator Author

Superseded by #574, which publishes the same commit from a branch in this repository so CI and automated review can run.

This branch had an error being deployed

1 failed deployment
azure-prod — 4819ba8d Deployed Nov 11, 2026 by aakash-saravanan-db via e2e-test (26.12.1, 26_12_1) #1837
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

1 participant