Skip to content

migration repair fails over Supavisor Session Pooler while migration list succeeds #6393

Description

@SLMK14

Environment

  • macOS arm64
  • Supabase CLI 2.116.0 and 2.84.2
  • Direct database IPv6 connectivity is unavailable
  • Supavisor Session Pooler on port 5432 is used as the IPv4 connection path

Scenario

The database schema is already correctly applied. The only required operation is reconciling migration history with the supported history-only command.

  1. Confirm that supabase migration list works through the Session Pooler connection.
  2. Run supabase migration repair <version> --status applied through the same connection.
  3. Run supabase db push --dry-run; it detects the expected local/remote migration-history mismatch.

Actual behavior

  • CLI 2.116.0 fails with LegacyDbExecError / effect/sql/SqlError: Connection error.
  • CLI 2.84.2 connects and then hangs or times out while updating migration history.
  • Migration list and other read operations continue to work through the same Session Pooler connection.
  • Read-back confirms that migration history remains unchanged after the failed repair attempts.

Expected behavior

migration repair --status applied should complete through the supported Supavisor Session Pooler connection and update migration history only. It should not require reapplying the already-correct schema.

Metadata

Metadata

Assignees

No one assigned

    Type

    No type

    Projects

    No projects

    Milestone

    No milestone

    Relationships

    None yet

    Development

    No branches or pull requests

    Issue actions