How to run a model in the background as a persistent service

What this does

Converts the Ollama server into a background service that remains available across terminal sessions, reboots, and logout events. After this guide the model API will be accessible without manual intervention.

Steps

1. Launch the Ollama server with nohup backgrounding. Detaches the process from the controlling terminal so it persists after logout.

   nohup ollama serve > /tmp/ollama.log 2>&1 &
   

   Expected output: A process ID printed by the shell.

2. Verify the server is listening. Confirms the server is accepting connections on port 11434.

   ss -tlnp | grep 11434
   

   Expected output: A line showing LISTEN state on port 11434 with the Ollama PID.

3. Set up a systemd service for automatic startup. Creates a unit file that survives reboots.

   sudo tee /etc/systemd/system/ollama.service << 'EOF'
   [Unit]
   Description=Ollama Server
   After=network-online.target
   Wants=network-online.target
   

[Service] ExecStart=/usr/local/bin/ollama serve Environment=OLLAMA_HOST=0.0.0.0 Restart=always RestartSec=3

[Install] WantedBy=multi-user.target EOF

4. **Enable and start the service.** Reloads systemd configuration, enables auto-start, and launches the server.
```bash
sudo systemctl daemon-reload && sudo systemctl enable ollama && sudo systemctl start ollama

Expected output: No errors; systemctl status shows "active (running)".

Verification

curl http://localhost:11434/api/tags
# Expected: JSON array containing model entries with "name" and "size" fields

Common failures

• address already in use - Port 11434 is occupied by another process; kill it with fuser -k 11434/tcp.
• Environment=OLLAMA_HOST ignored - Variable placed in wrong systemd section; must be under [Service].
• permission denied - User lacks write access to the log directory; use /tmp/ollama.log or adjust permissions.
• service starts then dies - ExecStart path is incorrect; verify with which ollama.

Operator checkpoint

Before treating this as solved, write down the local runtime, model or package version, hardware/backend if relevant, and the verification output. This keeps the guide useful as a Will-It-Run style decision instead of a one-off command transcript.

Related guides

• How to configure the Ollama serve port for network access
• How to stop a running model and free up memory