An interesting addition: I tried to provide the IP of our netmaker server directly instead of using the server name (DNS resolution seemed to be problematic potentially), and I got a different error now:

./netclient join -k <key> -n <network> --apiserver api.<ip>:443 --grpcserver grpc.<ip>:443
[netclient] joining <network> at grpc.<ip>:443
[netclient] error installing: rpc error: code = Unavailable desc = connection error: desc = "transport: Error while dialing dial tcp: lookup grpc.<ip> on 127.0.0.1:53: no such host"
Failed to authenticate: rpc error: code = Unauthenticated desc = Something went wrong with Auto Login: rpc error: code = Unavailable desc = connection error: desc = "transport: Error while dialing dial tcp: lookup grpc.<ip> on 127.0.0.1:53: no such host"
[netclient] removed <network> network locally
[netclient] removed systemd remnants if any existed
[netclient] removed systemd remnants if any existed

I am still kind of suspicious of the firewall though

@User , Why aren't you joining with the token instead of key and specifying api and grpc server address? 🙂

you have to specify grpcssl as well

The token length breaks OpenWRT's terminal so I'm using

token-convert.sh

to work around it 🙂

Oh, is that another flag? What should I specify there?

on .... also you don't need the --apiserver

"Tells clients to use SSL to connect to GRPC if 'on'. Disable if 'off'. Off by default.",

Cool! I'm not on my computer anymore but I'll try it out tomorrow. Thanks!

Huh...Fired up a droplet on digitalocean, got netmaker up and running with nm-quick.sh in a minute and it just works 😛

is it that surprising? 😂

Didn't imagine it would be that easy 😆

I might have been a bit too quick to celebrate though...I see now that the node is having the warning status...but this time the node has a peer for the comms interface 🤔

I managed to get my router onto Netmaker by the way! Got it working after upgrading from v0.9.4 to v0.12.0 and using the

./netclient join -t

rather than the

./netclient join -k ...

command. Seems like the token in v0.12.0 is much shorter than in the older version, which made it possible to join using the token 😄

As a side note, the command generated from

token-convert.sh

does not work in v0.12.0, although, it may not be necessary to use it since the token is shorter

Good point, we need to update that script and the docs, since the token is useable now. Want to open a git issue for this?

Now that Bitsauce and I have migrated to 0.12.0 we've noted an automatically generated network with a random name already created, with the description "This is your comms network". We found no note of it in the docs. 🤔 What is this network intended for? Additionally, lots of really cool features added 🎉 so great work thus far, looking very much forward to 1.0!

the comms network was introduced in 0.11.0. "Comms network - Every node now joins the "comms network" before joining any other network. This greatly increases the stability of client-server communications by having a separate network for updates. "

Aha, I see. That makes sense, thanks for the quick reply

Np 🙂 The documentation is a bit scattered.

And now I see that one of my nodes actually keeps a healthy status...progress 😄

Yeah we should put in some docs for the comms network. Long term it should help with reliability but short term, it’s got some issues

General networking, I set up a default network and added two nodes. They can ping the coordination server, but not each other. Is there a firewall setting or something I need to enable? The nodes are on the same LAN (for now) as well. I tried the Relay option on the server which separates the nodes and also didn't allow pings through. I tried setting the server as egress or ingress, and that also didn't work. Edit: It wasn't working, but with the server as ingress node eventually (minutes later?) pings were going through. I stopped and restarted and now it seems hung. And traffic was routing through the server which isn't ideal.

3934 packets transmitted, 36 received, +3 errors, 99.0849% packet loss, time 4026493ms

@User I really like Netmaker but would love to get an answer to this question anyway: If one enables the optional wireguard setting for the flannel-backend in K3s is there any benefit of using Netmaker over this solution?

i do not know if i am missing something here but last week I cleared my old 0.9.4 install and removed all the docker volumes and what not. Changed the version number and ran a docker up. I had to recreate networks, and the expected things. But there is something strange where if i go to nodes. the netmaker server says it is running v0.11.1. But on the left it says it is running server and UI v0.12.0

how did you install the clients?

i did not install a client on the server, as in I am just using the docker compose file to spin up netmaker

everything works fine on linux OS's

i can ping different nodes etc (one issue with OS X client, not had time to look what the issue is yet). I just dont know if it is a UI bug or is something running in the netmaker container that mentions v0.11.1

UI: v0.12.0
Server: v0.12.0

&

10.20.30.5    v0.12.0    HEALTHY
10.20.30.3    v0.12.0    HEALTHY
10.20.30.254    v0.11.1    HEALTHY <-- This is the netmaker server
10.20.30.2    v0.12.0    WARNING
10.20.30.4    v0.12.0    HEALTHY
10.20.30.1    v0.12.0 HEALTHY

might be a browser cache ---> try refreshing the page