Update start.sh

start.sh

@@ -4,13 +4,19 @@ set -e
 echo "🚀 Starting Ollama + FastAPI Server..."
 
 # Set Ollama environment variables
+export OLLAMA_HOME=/app/.ollama
 export OLLAMA_HOST=0.0.0.0:11434
 export OLLAMA_ORIGINS="*"
 export OLLAMA_KEEP_ALIVE=5m
 
-# Start Ollama in background (as root - this is fine in containers)
+
+# Create Ollama directory if it doesn't exist
+mkdir -p /app/.ollama
+chmod 755 /app/.ollama
+
+# Start Ollama in background
 echo "📡 Starting Ollama service..."
-ollama serve &
+nohup ollama serve > /app/ollama.log 2>&1 &
 OLLAMA_PID=$!
 
 # Wait for Ollama to be ready
@@ -24,26 +30,19 @@ for i in {1..60}; do  # Increased timeout to 60 attempts
     sleep 2
 done
 
-# Check if Ollama is actually running
-if ! curl -f http://127.0.0.1:11434/api/tags >/dev/null 2>&1; then
-    echo "❌ Ollama failed to start properly"
-    echo "🔍 Checking Ollama process..."
-    ps aux | grep ollama || echo "No Ollama process found"
+
+# Check if Ollama started successfully
+if curl -f http://127.0.0.1:11434/api/tags >/dev/null 2>&1; then
+    echo "✅ Ollama is running!"
     
-    echo "🔄 Attempting to restart Ollama..."
-    killall ollama 2>/dev/null || true
-    sleep 2
-    ollama serve &
-    sleep 10
+    # Try to pull model in background (don't block startup)
+    echo "📥 Starting model download in background..."
+    nohup ollama pull llama3.2:1b > /app/pull.log 2>&1 &
+    echo "📝 Model download started - check /app/pull.log for progress"
+else
+    echo "⚠️ Ollama not responding - check /app/ollama.log for details"
 fi
 
-# Try to pull the model (no user switching)
-echo "📥 Pulling Llama 3.2 1B model..."
-ollama pull hf.co/bartowski/Llama-3.2-1B-Instruct-GGUF:Q4_K_M || {
-    echo "⚠️ Model pull failed, trying smaller model..."
-    ollama pull llama3.2:1b || echo "⚠️ Fallback model pull also failed, continuing..."
-}
-
 echo "🎉 Setup complete!"
 echo "📡 Ollama API: http://localhost:11434"
 echo "📖 FastAPI: http://localhost:7860"