Deployment Guide¶
If you want to deploy MoLab in a production environment, you can follow the instructions below to set up the backend and worker components on a server. How you distribute the clients is up to you.
Option 1: Bare Metal¶
Let's assume we want to run the worker and backend on the same server for now, which is the simplest setup, because we can just follow the regular installation guide. In a second step we'll add more workers on different servers.
On the first server, clone the repository and make sure you have the necessary dependencies installed (uv and just). Then setup the Python environment and download the model weights:
You can then run the backend and worker components on the server:
By default, the backend server will listen on 0.0.0.0:8000, which means it will accept connections from any IP address.
Adding More Workers¶
Now, let's add another worker on a different server. Make sure you have the repository cloned and setup as before.
By default, a worker will connect to the backend server running on localhost:8000.
If you want to change this behavior, you can set the MOLAB_GATEWAY_HOST and MOLAB_GATEWAY_PORT environment variables accordingly:
Then you can start the worker:
You should see a message that the worker has connected to the backend server, which means it is ready to accept inference requests.
For easier deployment, you can also use Docker, which we'll cover in the next section.
Option 2: Docker¶
Each component has a Dockerfile in its root directory, which follows a similar setup as the bare metal setup, but with the added benefit of containerization.
Gateway Image¶
To build the backend image, you can run the following command:
Then export the image to a tar file and transfer it to your server:
On the server, load the image and start the container:
This will start the backend server and expose it on port 8000.
Worker Image¶
If you haven't done so already, make download the model weights:
Now you can follow the same steps as above, but for the worker component:
Then export the image to a tar file and transfer it to your server:
On the server, load the image and start the container, this time with the necessary environment variables pointing to the backend server:
docker load -i molab_worker.tar
docker run molab_worker \
-e MOLAB_GATEWAY_HOST=your-host \
-e MOLAB_GATEWAY_PORT=your-port
docker-compose¶
As an alternative approach to the steps outline above, you can also use docker-compose.
There is a docker-compose.yml file in the root directory as example that you can use to start the backend and two worker instances locally.
It specifies the necessary environment variables for the backend and workers, as well as the GPU requirement for the workers.