Adding node to Libcluster not working

Questions / Help
1

I currently have an Elixir App running in EC2. It runs on 3 Intel Instances and it is a Phoenix App fwiw.

I’m attempting to add a 4th server which is ARM based and the newer AL2023 distribution. The issue is that the new server cannot connect to the existing cluster. The AWS guy says that all ports are open between the 4 servers, so I need some assistance troubleshooting this problem in case there are networking issues I need to have the admin address.

I’ve run the cluster locally between an Intel Macbook and an ARM based Macbook and the 2 machines were able to cluster just fine, so I don’t think it’s an Architecture issue. I’m not sure how to troubleshoot this beyond trying a Node.ping(:app@server) which yields a :pang response or a Node.connect(:app@server) which yields a false response.

Using telnet, it appears that port 4369 is open between the servers.

Any help is appreciated :slight_smile:

2

Is the cookie on the new machine set to the same value as on the cluster?
What kind of strategy are you using for libcluster? Epmd?

2 Likes
3

It was the cookie. I was releasing in two steps, so two cookies were being generated, one for the Intel instances and one for the ARM instance.

Thanks!

2 Likes