You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
$ ./llama.cpp/build/bin/llama-cli --version
version: 4336 (08ea539)
built with cc (GCC) 14.2.1 20240910 for x86_64-pc-linux-gnu
Operating systems
Linux
Which llama.cpp modules do you know to be affected?
llama-server
Problem description & steps to reproduce
I am running llama-server behind Caddy as an authenticating reverse proxy. Once authenticated, all requests should keep sending the "Authorization: Basic TOKEN" header. This was the case before commit 235f6e1.
Since commit 235f6e1, that Authorization header is only sent for the initial connection to "/" and not for subsequent completion requests. Instead, it gets overwritten with "Authorization: undefined".
The symptom is that all completion requests fail. The never reach the server. See the screenshots below. The main difference is at the bottom right under "Request headers" (Authorization).
I understand that 235f6e1 uses a new kind of completion requests which I'm not too familiar with. Please let me know if there's any further information I can provide to help diagnose this issue.