yt-claw-v9 / start.sh
Arshit Malik
fix: runtime nginx config with 600s timeout + direct Ollama proxy
9e1dd7a
Raw
History Blame Contribute Delete
1.59 kB
#!/bin/bash
set -e
pkill -f "health" 2>/dev/null || true
crontab -r 2>/dev/null || true
echo "[boot] Writing custom nginx config..."
cat > /etc/nginx/sites-available/default << 'NGX'
server {
listen 7860;
location /ollama/ {
rewrite ^/ollama/(.*) /$1 break;
proxy_pass http://127.0.0.1:11434;
proxy_http_version 1.1;
proxy_set_header Host $host;
proxy_read_timeout 600s;
proxy_send_timeout 600s;
}
location / {
proxy_pass http://127.0.0.1:8080;
proxy_set_header Host $host;
proxy_read_timeout 600s;
proxy_send_timeout 600s;
}
}
NGX
echo "[boot] Nginx config written"
echo "[boot] Starting health+chat server on :8080..."
python3 /app/health_server.py &
echo "[boot] Health server PID: $!"
echo "[boot] Starting nginx..."
nginx
echo "[boot] Starting Ollama..."
export OLLAMA_FLASH_ATTENTION=1
export OLLAMA_KV_CACHE_TYPE=q4_0
export OLLAMA_NUM_CTX=131072
OLLAMA_HOST=127.0.0.1 OLLAMA_NUM_PARALLEL=1 OLLAMA_MAX_LOADED_MODELS=1 ollama serve &
for i in $(seq 1 30); do
if curl -sf http://127.0.0.1:11434/api/tags > /dev/null 2>&1; then
echo "[boot] Ollama ready after ${i}s"; break
fi
sleep 2
done
echo "[boot] Pulling IQ3_XS model for 128K context..."
ollama pull hf.co/bartowski/DeepSeek-R1-Distill-Qwen-14B-GGUF:IQ3_XS &
PULL_PID=$!
wait $PULL_PID
echo "[boot] Model ready"
echo ""
echo "=============================================="
echo " YT Claw Chat LIVE - 14B IQ3_XS (128K)"
echo "=============================================="
echo ""
sleep infinity