This message was deleted.

I have some special networking where I know traffic destined for

127.0.0.1

is not sourced normally - my machines are multi-homed and traffic bound for 127.0.0.1 is sourced from a VLAN interface. I think this is what's causing the issue, is there a way to force this connectivity to use the

node-external-ip

or

node-ip

?

Connections to the local etcd instance always use the loopback address. And that is just how etcd formats it's address internally, it's an implementation detail that you don't need to worry about.

I'm seeing this from the first server - I guess it's possible I need to bring the next 2 online perhaps? Not sure - I ran through the "hello world" for etcd and understand a little more about the configuration now.

wait are you using your own external etcd? or are you using the embedded one? You shouldn’t need to do anything with the embedded one, just make sure the ports are open and the server nodes can reach each other at their internal IPs.

I have not been, I've never setup etcd so I installed it separately to get a better understanding of how it works, since it's its own clustered KV store 😕

there are SO many options for doing cluster membership auto-discovery but it turns out most folks just start with one and then add more by hand and then never touch it again lol