Can't access llama-server from outside although I have exposed ports. #9079
Replies: 4 comments 2 replies
-
Install nginx as reverse proxy. Here is a (simple) example config:
|
Beta Was this translation helpful? Give feedback.
-
@opendeluxe what happen when you try to curl host_system_ip:8000 from the host system? Does it freeze too? |
Beta Was this translation helpful? Give feedback.
-
I'd recommend against exposing llama-server publicly for security reasons. Instead, forward the remote port to your local machine using ssh:
Then you should be able to access http://localhost:8000 in your browser. |
Beta Was this translation helpful? Give feedback.
-
The problem you're mentioning is not related to llama.cpp. Most VPS provider block all ports by defualt, you need to allow some of them manually. For hetzner: https://docs.hetzner.com/konsoleh/account-management/configuration/openports/ |
Beta Was this translation helpful? Give feedback.
-
I started
llama-server
on a Hetzner server with the following command:docker run -p 8000:8000 --gpus all -v ./models:/models ghcr.io/ggerganov/llama.cpp:server-cuda -m ... --n-gpu-layers 99 --ctx_size 1024 --host 0.0.0.0 --port 8000
In the terminal of the host system, I can curl it: curl localhost:8000
It prints html and javascript to the console.
Shouldn't I be also able to access this application from outside my server via web browser? I have exposed the ports via docker using the switch "-p 8000:8000".
When I make a request to the IP-adress of the server and the corresponding port 8000, it just hangs, eventually timing out later.
I even can not see any new log entry of llama-server in that time.
I don't run any firewall.
ufw status
showsinactive
.Beta Was this translation helpful? Give feedback.
All reactions