Originally created by @jackrosenberg on GitHub (Jul 30, 2025).
Hello,
I've upgraded Pangolin from 1.2.0 to 1.7.3, and am unable to access resources anymore. I thought this could be a migration error or something so I did a clean install, which has not resolved the error. Newt connects normally, and there's nothing unusual in Traefik, Gerbil or Pangolin's logs. Anyone got any idea how i can debug this?
Originally created by @jackrosenberg on GitHub (Jul 30, 2025).
Hello,
I've upgraded Pangolin from 1.2.0 to 1.7.3, and am unable to access resources anymore. I thought this could be a migration error or something so I did a clean install, which has not resolved the error. Newt connects normally, and there's nothing unusual in Traefik, Gerbil or Pangolin's logs. Anyone got any idea how i can debug this?
@jackrosenberg commented on GitHub (Jul 30, 2025):
curl localhost:3001/api/v1/traefik-config gives the correct config, but curl ocalhost:3001/api/v1/gerbil/get-config gives: {"data":null,"success":false,"error":true,"message":"The requests url is not found - /api/v1/gerbil/get-config","status":404,"stack":null}
@jackrosenberg commented on GitHub (Jul 30, 2025):
`curl localhost:3001/api/v1/traefik-config` gives the correct config, but
`curl ocalhost:3001/api/v1/gerbil/get-config` gives:
`{"data":null,"success":false,"error":true,"message":"The requests url is not found - /api/v1/gerbil/get-config","status":404,"stack":null}`
@oschwartz10612 commented on GitHub (Jul 31, 2025):
Interesting. Sorry you are having this issues. We just released 1.8.0 yesterday with some fixes. Could you try that?
@oschwartz10612 commented on GitHub (Jul 31, 2025):
Interesting. Sorry you are having this issues. We just released 1.8.0 yesterday with some fixes. Could you try that?
Found it, looks like I removed a bit too much from the config.yml following the 1.5.0 updates. My bad! (server.internal_hostname = "localhost")
@jackrosenberg commented on GitHub (Aug 1, 2025):
Found it, looks like I removed a bit too much from the `config.yml` following the 1.5.0 updates. My bad! (`server.internal_hostname = "localhost"`)
I mean, it might still be useful to investigate why server.internal_hostname doesn't default to localhost. It defaults to pangolin: https://github.com/fosrl/pangolin/blob/92e69f561f71f9214aba795226f16b78a9a10255/server/lib/readConfigFile.ts#L67
...which means Pangolin will break if it's not running in a machine called pangolin. This isn't an issue in containers, I guess, but it's still a really weird decision. localhost means it'll work regardless of the hostname, anyway.
@SigmaSquadron commented on GitHub (Aug 1, 2025):
I mean, it might still be useful to investigate why `server.internal_hostname` doesn't default to `localhost`. It defaults to `pangolin`: https://github.com/fosrl/pangolin/blob/92e69f561f71f9214aba795226f16b78a9a10255/server/lib/readConfigFile.ts#L67
...which means Pangolin will break if it's not running in a machine called `pangolin`. This isn't an issue in containers, I guess, but it's still a really weird decision. `localhost` means it'll work regardless of the hostname, anyway.
Pangolin doesn't create interfaces, I don't think?
internal_hostname is the host pangolin communicates with. In most cases, that'll be localhost.
@SigmaSquadron commented on GitHub (Aug 1, 2025):
Pangolin doesn't create interfaces, I don't think?
`internal_hostname` is the host pangolin communicates with. In most cases, that'll be `localhost`.
@oschwartz10612 commented on GitHub (Aug 6, 2025):
This is actually used to tell badger in the traefik config how to reach
pangolin. The plugin pulls it from the config values sent in the traefik
config http provider.
So in this case pangolin makes sense because it is the best default for
the containers I think because the traefik container reaches pangolin at
that hostname. But you are correct that when not containerized it should
be something like localhost.
I think container deployment is by far the most common but I know some
folks are working on a NIXOS deal!
@oschwartz10612 commented on GitHub (Aug 6, 2025):
This is actually used to tell badger in the traefik config how to reach
pangolin. The plugin pulls it from the config values sent in the traefik
config http provider.
So in this case pangolin makes sense because it is the best default for
the containers I think because the traefik container reaches pangolin at
that hostname. But you are correct that when not containerized it should
be something like localhost.
I think container deployment is by far the most common but I know some
folks are working on a NIXOS deal!
I think container deployment is by far the most common but I know some
folks are working on a NIXOS deal!
We are the folks ;) Its pretty much done, we've been waiting on upstream merges for a while now. Will let you know when all is done!
@jackrosenberg commented on GitHub (Aug 6, 2025):
> I think container deployment is by far the most common but I know some
> folks are working on a NIXOS deal!
We are the folks ;) Its pretty much done, we've been waiting on upstream merges for a while now. Will let you know when all is done!
@oschwartz10612 commented on GitHub (Aug 6, 2025):
Ahh sorry haha not keeping track. Thanks so much for doing this we really appreciate the contribution! I hope nix crowd love it :}
@oschwartz10612 commented on GitHub (Aug 6, 2025):
Ahh sorry haha not keeping track. Thanks so much for doing this we really appreciate the contribution! I hope nix crowd love it :}
Blocking a user prevents them from interacting with repositories, such as opening or commenting on pull requests or issues. Learn more about blocking a user.
Originally created by @jackrosenberg on GitHub (Jul 30, 2025).
Hello,
I've upgraded Pangolin from 1.2.0 to 1.7.3, and am unable to access resources anymore. I thought this could be a migration error or something so I did a clean install, which has not resolved the error. Newt connects normally, and there's nothing unusual in Traefik, Gerbil or Pangolin's logs. Anyone got any idea how i can debug this?
@jackrosenberg commented on GitHub (Jul 30, 2025):
curl localhost:3001/api/v1/traefik-configgives the correct config, butcurl ocalhost:3001/api/v1/gerbil/get-configgives:{"data":null,"success":false,"error":true,"message":"The requests url is not found - /api/v1/gerbil/get-config","status":404,"stack":null}@oschwartz10612 commented on GitHub (Jul 31, 2025):
Interesting. Sorry you are having this issues. We just released 1.8.0 yesterday with some fixes. Could you try that?
@jackrosenberg commented on GitHub (Aug 1, 2025):
Found it, looks like I removed a bit too much from the
config.ymlfollowing the 1.5.0 updates. My bad! (server.internal_hostname = "localhost")@SigmaSquadron commented on GitHub (Aug 1, 2025):
I mean, it might still be useful to investigate why
server.internal_hostnamedoesn't default tolocalhost. It defaults topangolin: https://github.com/fosrl/pangolin/blob/92e69f561f71f9214aba795226f16b78a9a10255/server/lib/readConfigFile.ts#L67...which means Pangolin will break if it's not running in a machine called
pangolin. This isn't an issue in containers, I guess, but it's still a really weird decision.localhostmeans it'll work regardless of the hostname, anyway.@jackrosenberg commented on GitHub (Aug 1, 2025):
IFAK, internal_hostname != hostname. It's the name of the network interface the pangolin stack uses
@SigmaSquadron commented on GitHub (Aug 1, 2025):
Pangolin doesn't create interfaces, I don't think?
internal_hostnameis the host pangolin communicates with. In most cases, that'll belocalhost.@jackrosenberg commented on GitHub (Aug 2, 2025):
@oschwartz10612, could you explain this to us?
@oschwartz10612 commented on GitHub (Aug 6, 2025):
This is actually used to tell badger in the traefik config how to reach
pangolin. The plugin pulls it from the config values sent in the traefik
config http provider.
So in this case pangolin makes sense because it is the best default for
the containers I think because the traefik container reaches pangolin at
that hostname. But you are correct that when not containerized it should
be something like localhost.
I think container deployment is by far the most common but I know some
folks are working on a NIXOS deal!
@jackrosenberg commented on GitHub (Aug 6, 2025):
We are the folks ;) Its pretty much done, we've been waiting on upstream merges for a while now. Will let you know when all is done!
@oschwartz10612 commented on GitHub (Aug 6, 2025):
Ahh sorry haha not keeping track. Thanks so much for doing this we really appreciate the contribution! I hope nix crowd love it :}