The error “curl: (7) Failed to connect to localhost port 9000” means the vLLM server is not reachable at that address/port, even though your terminal shows the server command was executed. This can happen if the server failed to start, is still loading the model (which can take time), or crashed due to an error. You should check the full server logs for errors or messages indicating successful startup (look for “Uvicorn running on…” or “Application startup complete”) before sending requests. Also, ensure no firewall or network restrictions block access to port 9000, and that you are using the correct host/port combination for your environment (e.g., WSL, Docker, or bare metal) as networking can differ in these setups.
If you are running in WSL or a similar environment, networking issues are common—sometimes changing WSL’s network mode or setting a static DNS can help. Also, make sure the model path is correct and the model files are accessible. For more troubleshooting steps and common solutions, see this issue and this issue.
Would you like a step-by-step troubleshooting guide?
Sources:
Hint: Mention @RunLLM in the post for followups.
