In a full app, migrations combine system catalogs and module-owned scripts.
bun run db:setup
bun run db:migrate
bun run db:statusUse db:setup for a new development database. It creates the configured
database when missing and applies every migration. Use db:migrate for an
existing database.
Production sequence
- Build the application image.
- Run
bun run db:migrateonce as a release job. - Start API and worker processes only after that job succeeds.
Applied migrations are tracked by owner and migration ID in
damat._damat_migration_logs, making repeated status and migrate commands
idempotent. All framework-owned durability, job, event, and pipeline relations
live in the dedicated damat schema; application, module, and link tables stay
in their configured application schemas.
The 1.0.6 migration preflight moves an existing public migration tracker before
reading its history. Forward system migrations relocate the remaining legacy
public._damat_* tables with ALTER TABLE ... SET SCHEMA, preserving data,
indexes, constraints, identities, grants, and foreign keys. Never start 1.0.6
API or worker processes against a database that has not completed this step.
Creating a missing database requires CREATEDB. Creating the damat schema
requires database CREATE; migrating an already prepared database does not
require an administrative or superuser role.