DistributedSystems-CS

1. Build and start the system

From the repository root, run:

docker compose up --build

This command will:

• Build the shared image from the Dockerfile.

• Start:

  • four file_service_* containers,
  • one coordinator container,
  • one worker container - scalable to more.

You can view logs with:

docker compose logs -f

---

2. Scaling the workers

You can simulate multiple workers by scaling the worker service:

docker compose up --build --scale worker=5

This will start five independent worker containers (named worker_1, worker_2, etc.), each connecting to the same coordinator and file service.

---

Environment Configuration

The containers communicate via a shared Docker network. Each component reads its configuration from environment variables (set in docker-compose.yml):

Variable	Default	Description
FILE_SERVICE_HOST	file_service	Hostname of the main file service container
FILE_SERVICE_PORT	50051	It's default port
COORDINATOR_HOST	coordinator	Hostname of the coordinator container
COORDINATOR_PORT	50052	Coordinator gRPC port

In local (non-Docker) runs, these default to localhost and can be overridden with --env.

---

Code Overview

• services/file_service/server.py: gRPC file server implementation.
• services/coordinator/coordinator.py: Coordinator gRPC service.
• services/coordinator/worker.py: Worker logic connecting to coordinator and file service.
• proto/generated/: Generated protobuf and gRPC Python files.

All processes are started using Python’s module syntax (e.g. python3 -m services.coordinator.coordinator).

---

Local Development

If you’re actively editing the source code, you can mount your local directory into the containers by adding this under each service in docker-compose.yml:

volumes:
  - .:/app

Then code changes will be reflected without rebuilding the image.