I’m using docker, I get this in the logs if I stop the server:
Failed to create client. Ensure the server is running. URL: https:///api/clients/?client-name=dockerclient&user=dockerclient
the web admin ui is up and working, and if I browse to that Ip I get “Invalid method for /api/clients/” also it is passing the user and client-name as query string params
The URL is correct with my FQDN name in it, something in the formatting that I used broke it, but the url root is fine.
Hey @patrickrobbins, I’m not sure exactly what problem you’re having. You said “I get this in the logs if I stop the server”. Why are you stopping the server? Maybe this is a typo?
What happens if you try creating a tunnel?
If I don’t stop the container, it seems to just peg trying to retry connecting.
I know you don’t use docker much, I tried to use the docker client example from github + the usage instructions.
services:
boringproxy:
container_name: boringproxy
image: jangwonseok/boringproxy:latest
restart: unless-stopped
command: [“client”, “-server”, “boring.fqdnroot.com”, “-token”, “token”, “-user”, “userFromWebUi”, “-client-name”, “clientnameFromWebUi”, “-cert-dir”, “/certmagic”]
Is it expected that the client would call the server with the /api/clients/ endpoint?
Can you link to the docker client example you’re following? I don’t use docker for boringproxy personally.
Also, can you trying running with just -server and -token? -user and -client-name are usually redundant with current versions of boringproxy.
I found some YML docker compose files in the github repo. I will try with only -server and -token. If there is no client name, how will it know the client if mutliple use the same token? ( I have a separate user and token for this first endpoint, so maybe that will be ok for me)
Yes that is expected
For security you should use separate tokens for each client anyway, but yes in that case you’ll want to pass the client ID. Honestly I’m not even sure if that would work.