0

I have a Erlang cluster with some nodes. One day, I want to make a new node to join to the cluster by using net_adm:ping to ping to a specific node. But I got failure. Surely that this is not a cookie problem, or firewall neither because the cluster is working well. Just only connection from the specific node in the cluster to a new node got failure. So, I think this is an EPMD issue (maybe), then I want to reset the EPMD on the specific node. So anyone help me to point out how to reset EPMD without reset VM in order to a new node can connect to successfully?

Thanks so much!!!

Duke
  • 21
  • 7
  • There isn't any way to reset epmd itself, but there are things your can do from Erlang nodes. [This article](https://habr.com/en/post/49535/) describes what to do in this (or a similar?) scenario. It's in Russian, but Google Translate does a decent job. – legoscia Jun 21 '19 at 10:53
  • 1
    This article I found before, then worked follow. However, it didn't work for me. Maybe I was wrong something. I did as bellow: 1. In the EPMD damaged node, I forced kill EPMD and restart it in daemon mode 2. From one node in the cluster (because this node is working well with the damaged one), I used rpc call to register_node in the damaged one as the article guided. 3. Used net_adm ping from a new node to the damaged node, but got unlucky. PS: The new node can connect to other nodes in the cluster without any problem. Thanks so much! – Duke Jun 22 '19 at 10:53
  • 1
    I'd like update some more information, that is I couldn't find out SEND_CHALLENGE and SEND_CHALLENGE_REPLY messages between 2 nodes - the new node and the damaged epmd node while I was using net_adm ping command. So is there anybody has ever got such kind of this issue? – Duke Jun 25 '19 at 04:35

0 Answers0