'Could not connect to'-Error when using `remote` to connect to running elixir release

Questions / Help
, , ,
1

Hey there,

after managing to get my little side project running and deployed I wanted to try out BEAM’s inspection features. But somehow BEAM does not find the runnig instance on the host

$ ./bin/tilemirror remote
Could not connect to "tilemirror"

The release is running on a shared host and is started using the following supervisor script

[program:tilemirror]
directory=~/tilemirror/
command=~/tilemirror/bin/tilemirror start
startsecs=10

tilemirror/ is the release folder

ls tilemirror/
bin  erts-14.1.1  erts-14.2.1  lib  releases

and it’s working as intended

$ netstat -tulpn | grep 4000
tcp        0      0 0.0.0.0:4000            0.0.0.0:*               LISTEN      11498/beam.smp

I tried to set the REMOTE_NODE variable explicitly, but it did not make a difference.

My understanding of the BEAM VM is way to shallow to have an idea what could be the problem. Is there an obvious reason?

2

Not clear for your case but I have suggestion.

./bin/tilemirror remote

As I know, this for run on same host (same container for k8s/docker) only. I use this for debug runtime in product that runs in container).

For join from other node you need join cluster by manual (or using libcluster) add need add rule to firewall for epmd port (if you use this).

3

Mh.. I’m on the same host as the running beam instance and there’s no docker involved. Maybe supervisord has something todo with it - process ownership or something like that.

I could check later, if it makes any difference, if I start it as a user. If not, I can only guess that it has something to do with the configuration of the shared host. But I don’t know where to start looking.

4

Might be a stupid question but do you have the same Erlang cookie set on both machines? :thinking: