Running the Stack

Operational reference for starting, managing, and troubleshooting the Docker Compose services.

This section details how to run, manage, and debug the core platform services locally.

Service Mapping & Ports

All core services run inside a shared Docker network bridge named nordstern-net. The table below lists the public and internal ports:

Service Name	Container Name	Host Port	Internal Port	Health Check Endpoint
db	`nordstern-platform-db`	`5432`	`5432`	Postgres ping (`pg_isready`)
secrets	`localstack`	`4566`	`4566`	LocalStack status `/health`
traefik	`traefik`	`80` (Web) `8090` (Dashboard)	`80` / `8080`	None
platform-api	`nordstern-platform-api`	`4000`	`4000`	HTTP GET `/health`
control-plane	`nordstern-control-plane`	`3002`	`3002`	HTTP GET `/health`
aggregator	`nordstern-aggregator`	`3005`	`3005`	HTTP GET `/health`
founder-console	`nordstern-founder-console`	`4001`	`3000`	HTTP GET `/`
admin-console	`nordstern-admin-console`	`4002`	`3000`	HTTP GET `/`

Always specify the generated env base file when starting services:

docker compose --env-file anchor-service/.env.base -f docker-compose.platform.yml up -d

To debug a crashing service (e.g. platform-api):

docker compose -f docker-compose.platform.yml logs -f platform-api

Tear down the containers without deleting databases:

docker compose -f docker-compose.platform.yml down

Tear down containers, networks, and remove the Postgres volume:

docker compose -f docker-compose.platform.yml down -v

Error: network nordstern-net declared as external, but could not be found or incorrect label.
Cause: The docker network was created with legacy configuration labels.
Fix: Delete the network manually and restart the stack:
```
docker network rm nordstern-net
```

Error: bind: address already in use on port 5432.
Cause: A local Postgres server is running on the host system.
Fix: Stop your host Postgres server or other running docker instances:
```
# On macOS
brew services stop postgresql
```

Error: Onboard redeem spinner hangs, stage stays stuck in "Funding accounts".
Cause: The Stellar Testnet is reset quarterly, or Friendbot is temporarily down.
Fix: Check network connection, retry the job via POST /api/v1/anchor-invitations/status/:jobId/retry.