Skip to main content

lf services

Manage LlamaFarm backend services (server, RAG worker, universal runtime).

Synopsis​

lf services <subcommand> [service-name] [flags]

Subcommands​

CommandDescription
statusCheck status of all services
startStart services
stopStop services

Available Services​

ServiceDescriptionDefault Port
serverMain FastAPI server8000
ragRAG/Celery workerN/A
universal-runtimeUniversal Runtime server for HuggingFace models11540

Orchestration Modes​

LlamaFarm supports multiple orchestration modes, controlled by the LF_ORCHESTRATION_MODE environment variable:

ModeDescription
native (default)Run services as native processes
dockerRun services as Docker containers
autoAuto-detect best mode (prefers native)

lf services status​

Check the current status of all LlamaFarm services without starting them.

lf services status [flags]

Flags​

FlagDescription
--jsonOutput status in JSON format

Output Information​

The status command shows:

  • Process/container running state
  • PID (native) or container ID (Docker)
  • Port mappings (Docker)
  • Health status (if service is running)
  • Log file location (native) or image information (Docker)
  • Uptime

Examples​

# Check status of all services
lf services status

# Get machine-readable JSON output
lf services status --json

Sample Output​

LlamaFarm Services Status
═════════════════════════

  Service             State      PID      Health    Uptime
  ──────────────────────────────────────────────────────────
  server              running    12345    healthy   2h 15m
  rag                 running    12346    healthy   2h 15m
  universal-runtime   stopped    -        -         -

lf services start​

Start LlamaFarm services.

lf services start [service-name] [flags]

Arguments​

  • service-name (optional): Specific service to start. If omitted, starts all services.

Examples​

# Start all services
lf services start

# Start only the main server
lf services start server

# Start only the RAG worker
lf services start rag

# Start only the Universal Runtime
lf services start universal-runtime

# Start all services using Docker
LF_ORCHESTRATION_MODE=docker lf services start

lf services stop​

Stop LlamaFarm services.

lf services stop [service-name] [flags]

Arguments​

  • service-name (optional): Specific service to stop. If omitted, stops all services.

Examples​

# Stop all services
lf services stop

# Stop only the main server
lf services stop server

# Stop only the RAG worker
lf services stop rag

# Stop all Docker containers
LF_ORCHESTRATION_MODE=docker lf services stop

Environment Variables​

VariableDescriptionDefault
LF_ORCHESTRATION_MODEOrchestration mode (native, docker, auto)native

Use Cases​

Development Workflow​

# Start all services for development
lf services start

# Check everything is running
lf services status

# When done, stop everything
lf services stop

Debugging Service Issues​

# Check which services are running
lf services status --json | jq '.services[] | select(.state == "running")'

# Restart a specific service
lf services stop server
lf services start server

Docker Deployment​

# Start services in Docker mode
LF_ORCHESTRATION_MODE=docker lf services start

# Check container status
lf services status

# Stop Docker containers
LF_ORCHESTRATION_MODE=docker lf services stop

Troubleshooting​

Service Won't Start​

  1. Check if the port is already in use:

    lsof -i :8000  # For server
    lsof -i :11540 # For universal-runtime

  2. Kill conflicting processes:

    kill <PID>

  3. Try starting the service again:

    lf services start server

Service Shows "unhealthy"​

  1. Check the service logs (native mode):

    # Log files are shown in status output
    tail -f ~/.llamafarm/logs/server.log

  2. Verify dependencies are running:

    • Server needs Redis for Celery
    • RAG worker needs ChromaDB
    • Universal Runtime needs sufficient memory for models

Docker Mode Issues​

  1. Ensure Docker is running:

    docker ps

  2. Check Docker container logs:

    docker logs llamafarm-server

See Also​