zerp/docs/PRODUCTION_DATABASE_CLEANUP.md

Production Database Cleanup Task

Purpose

Clean the production database so you can load new real data that will reflect across the system at all levels. This removes existing (e.g. test/demo) data and leaves the database in a state where:

• Schema and migrations are unchanged
• Base configuration is restored (pipelines, categories, departments, roles, default users)
• All business data (contacts, deals, quotes, projects, etc.) is removed so you can enter new real data

⚠️ Important

• Back up the production database before running any cleanup.
• Run this only on the production server (or a copy of it) when you are ready to wipe current data.
• After cleanup, the seed will recreate default login users (see credentials at the end of this doc). Change their passwords after first login if needed.

---

Option 1: Clean + Re-seed (Recommended)

This truncates all tables and then runs the seed so you get:

• Empty business data (contacts, deals, quotes, projects, inventory, etc.)
• Restored base data: departments, positions, permissions, employees, users, contact categories, product categories, pipelines, one warehouse

Steps on production server

Option A – One script (backup + clean + seed)

From the repo root on the production server:

bash backend/scripts/run-production-clean-and-seed.sh

This will prompt for YES, create a backup under backups/, then run clean-and-seed. Restart the app after.

Option B – Manual steps

1. Back up the database

   pg_dump $DATABASE_URL > backup_before_cleanup_$(date +%Y%m%d_%H%M%S).sql
   

2. Go to backend and run the cleanup script

   cd backend
   npm run db:clean-and-seed
   

3. Restart the application so it uses the cleaned DB.

4. Log in with one of the default users and start entering real data (contacts, deals, etc.).

---

Option 2: Full reset (drop and recreate database)

Use only if you want to drop the entire database and recreate it from migrations (e.g. to fix schema drift).

1. Back up the database (see above).

2. Reset and seed

   cd backend
   npx prisma migrate reset --force
   

   This will:

   • Drop the database
   • Recreate it from migrations
   • Run the seed

3. Restart the application.

---

What gets removed (Option 1)

All rows are removed from every table, including:

• Contacts, contact relationships, contact categories links
• Deals, quotes, cost sheets, contracts, invoices
• Projects, tasks, project members, expenses
• Inventory, products, warehouses, movements, assets
• Campaigns, activities, notes, attachments
• HR data: attendances, leaves, salaries, evaluations, etc.
• Audit logs, notifications, approvals
• Users, employees, departments, positions, permissions

Then the seed recreates only the base data (users, departments, positions, permissions, employees, contact/product categories, pipelines, one warehouse).

---

Default users after re-seed

Role	Email	Password	Access
General Manager	gm@atmata.com	Admin@123	Full system
Sales Manager	sales.manager@atmata.com	Admin@123	Contacts, CRM
Sales Representative	sales.rep@atmata.com	Admin@123	Basic CRM

Change these passwords after first login in production.

---

If you use Docker

Run the cleanup inside the backend container or against the same DATABASE_URL your app uses:

# Example: run inside backend container
docker compose exec backend npm run db:clean-and-seed

Or run the script from your host with DATABASE_URL set to the production DB connection string.

---

Troubleshooting

• "Cannot truncate because of foreign key" – The script uses CASCADE; if you see this, ensure you’re on PostgreSQL and using the provided script.
• Seed fails (e.g. duplicate key) – Ensure the cleanup step completed; run the script again (cleanup is idempotent, then seed runs once).
• App still shows old data – Restart the backend/API and clear browser cache; ensure DATABASE_URL points to the database you cleaned.