Deployment Commands Reference#

This reference provides quick lookup for common commands used to manage a running Dioptra deployment.

Note

All commands should be run from within your deployment directory (e.g., dioptra-deployment/).

See also

Update Your Deployment - Step-by-step guide for updating your deployment configuration.

Starting Dioptra#

Start all services in detached mode:

docker compose up -d

Install and start the systemd service (Linux only):

sudo cp ./systemd/dioptra.service /etc/systemd/system
sudo systemctl start dioptra
sudo systemctl enable dioptra  # Start on boot

Once started, access the frontend at http://localhost/ (or https://localhost/ if SSL/TLS is enabled).

Stopping Dioptra#

Stop all services (preserves data):

docker compose down

Stop and delete all data:

docker compose down -v

Warning

The -v flag deletes all named volumes, including your database and stored artifacts.

Stop the service:

sudo systemctl stop dioptra

Disable start on boot:

sudo systemctl disable dioptra

Checking Service Status#

View status of all containers:

docker compose ps

This shows running containers, health status, and exposed ports.

Viewing Logs#

All services:

docker compose logs

Specific service (e.g., REST API):

docker compose logs <deployment-name>-restapi

Follow logs in real-time:

docker compose logs -f

Press Ctrl+C to stop following.

Limit to recent entries:

docker compose logs --tail=100

Available service names:

Service names are prefixed with your deployment name. For example, if your deployment name is dioptra-deployment, the REST API service would be dioptra-deployment-restapi.

  • <deployment-name>-restapi - Dioptra REST API

  • <deployment-name>-nginx - NGINX reverse proxy

  • <deployment-name>-db - PostgreSQL database

  • <deployment-name>-dbadmin - pgAdmin4 database administration

  • <deployment-name>-redis - Redis queue

  • <deployment-name>-minio - Minio S3 storage

  • <deployment-name>-mlflow-tracking - MLflow tracking server

  • <deployment-name>-tfcpu-*, <deployment-name>-tfgpu-* - TensorFlow workers

  • <deployment-name>-pytorchcpu-*, <deployment-name>-pytorchgpu-* - PyTorch workers

Note

Replace <deployment-name> with the slugified version of your deployment name (e.g., dioptra-deployment for the default). Run docker compose ps to see the exact service names for your deployment.

Restarting Services#

Restart all services:

docker compose restart

Restart a specific service:

docker compose restart <deployment-name>-restapi

Restart all services:

sudo systemctl restart dioptra

See also