Anyone managed to connect mnesia on two machines with libcluster and gossip strategy?

Has anyone managed to connect mnesia on two machines with libcluster and gossip strategy?

2 Likes

I know it’s way too late but I just bumped at your question.

No, but I did manage to make it work using libcluster’s Kubernetes.DNS and LocalEpmd strategies (for prod and dev respectively), in case it helps somehow.

2 Likes

An input in local dev mode, the Gossip strategy works for me when I add a Process.sleep within the handle_info({:nodeup, node},..) in the Cache GenServer. This seems to be needed for the first start, when the tables are created on each node disc to overcome the communication latency.

1 Like

Still in a local cluster in dev monde. I modified my strategy and wrapped Mnesia into a supervised GenServer. I only have to loop on Mnesia start-up to ensure the successful launch of multiple nodes.