Ollama is a versatile AI model-serving tool designed for efficiency and ease of use. However, sometimes you need more control over your installation, like setting a custom directory. In this tutorial, we’ll cover step-by-step how to install Ollama in a custom location (/data/ollama) on your Linux server and configure it to be accessible remotely.
Step 1: Prepare the Custom Directory and User
First, create a dedicated directory and user for running Ollama securely:
sudo mkdir -p /data/ollama sudo useradd -r -s /bin/false -U -m -d /data/ollama ollama sudo chown -R ollama:ollama /data/ollama
Step 2: Download and Install Ollama Binaries
Download the Ollama Linux binary and extract it directly into your custom directory:
curl -L https://ollama.com/download/ollama-linux-amd64.tgz | sudo tar -xzf - -C /data/ollama sudo chown -R ollama:ollama /data/ollama
Step 3: Configure Ollama as a Systemd Service
To have Ollama run automatically and reliably, create a systemd service file:
[Unit] Description=Ollama Service After=network-online.target [Service] ExecStart=/data/ollama/bin/ollama serve WorkingDirectory=/data/ollama User=ollama Group=ollama Environment="HOME=/data/ollama" Environment="OLLAMA_MODELS=/data/ollama/models" Environment="OLLAMA_HOST=0.0.0.0:11434" Restart=always RestartSec=3 [Install] WantedBy=multi-user.target
Activate the Ollama service:
sudo mkdir -p /data/ollama/models sudo chown -R ollama:ollama /data/ollama/models sudo systemctl daemon-reload sudo systemctl enable --now ollama
Check the status to ensure everything works correctly:
sudo systemctl status ollama --no-pager
Step 4: Open the Firewall Port
Ensure your firewall allows incoming requests on Ollama’s default port (11434):
sudo firewall-cmd --permanent --add-port=11434/tcp sudo firewall-cmd --reload
Step 5: Verify the Installation
Test your installation locally with a quick command:
curl http://localhost:11434/api/version
You should see a JSON response indicating Ollama’s version.
Congratulations! You’ve successfully installed and configured Ollama at your custom location on Linux. Now your server is optimized for performance and flexibility.
