Skip to content

Flaky memory test: strict-drain row count races a shared change-log table #160

Open

Description

@tonibergholm-codento

snapshot_strict_drain_preserves_bad_row_on_block in crates/agentic-memory/tests/integration.rs failed once in CI (run 44734834443, PR #157) with left: 0, right: 1 at the post-snapshot row count, then passed on rerun.

The test plants a row in the shared public.agentic_change_log table through the admin pool. Tests that run in parallel can TRUNCATE that same table, so the planted row can vanish before the assertion. The other tests in this suite already run in their own database (#100); this one still uses the shared public table.

Fix: give this test its own database like the others, or plant the row in the per-test schema.

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Metadata

Metadata

Assignees

No one assigned

    Labels

    bugSomething isn't working

    Type

    No type

    Projects

    No projects

      Milestone

      No milestone

      Relationships

      None yet

      Development

      No branches or pull requests

      Issue actions