My ideal setup is with two networks: * One allowing access to private resources * One acting as an egress gateway for public internet traffic (for specific clients joined to this network only). The gateway is on a node separate from the netmaker server. The two networks should be as independent as possible. However, what I end up with is a setup where the netmaker server is joined to both networks, and its own traffic goes via the egress gateway.

I feel like the raw wg config would be to remove the egress allowed IPs from the peer on the netmaker server, but I'm not sure where I can do such a thing, since the server seems to be managed differently from clients

Some progress on my part...nodes are actually working and can ping each other now. But the nodes still fail to check in to the server as the status goes to warning->error. One node is on Windows and one Linux. Is there any logs that could give a clue?

netclient logs on Linux ->

journactl -fu netclient

netmaker logs

docker logs-f netmaker

netclient logs doesn't say much

-- Journal begins at Sat 2021-04-24 23:11:22 CEST. --
Mar 06 23:45:34 pop-os netclient[11103]: 2022/03/06 23:45:34 [netclient] checkin routine closed
Mar 06 23:45:34 pop-os netclient[11103]: 2022/03/06 23:45:34 [netclient] shutdown complete
Mar 06 23:45:34 pop-os systemd[1]: Stopping Netclient Daemon...
Mar 06 23:45:34 pop-os systemd[1]: netclient.service: Deactivated successfully.
Mar 06 23:45:34 pop-os systemd[1]: Stopped Netclient Daemon.
Mar 06 23:45:34 pop-os systemd[1]: Started Netclient Daemon.
Mar 06 23:45:34 pop-os netclient[11117]: 2022/03/06 23:45:34 [netclient] pulling latest config for Tnjg5NP0
Mar 06 23:45:34 pop-os netclient[11117]: 2022/03/06 23:45:34 [netclient] pulling latest config for default
Mar 06 23:45:34 pop-os netclient[11117]: 2022/03/06 23:45:34 [netclient] started comms network daemon, Tnjg5NP0
Mar 06 23:45:34 pop-os netclient[11117]: 2022/03/06 23:45:34 [netclient] netclient daemon started for network: Tnjg5NP0

netmaker docker logs throws a bunch of interesting errors

[netmaker] 2022-03-06 22:44:37 sent peer update for node Bernard on network: default
[netmaker] 2022-03-06 22:44:37 could not set peers on network default : file does not exist
[netmaker] 2022-03-06 22:44:37 ERROR RETRIEVING EXTERNAL PEERS ON SERVER: could not find any records
[netmaker] 2022-03-06 22:44:37 failed to parse interface
[netmaker] 2022-03-06 22:44:37 could not set peers on network default : file does not exist
[netmaker] 2022-03-06 22:44:37 ERROR RETRIEVING EXTERNAL PEERS ON SERVER: could not find any records
2022/03/06 22:45:31 [netclient] error running command: wg set nm-Tnjg5NP0 peer iCwwHbBXZcZ5De/+ulUwRhDi46PK1eTMR894ruzHbA4= persistent-keepalive 20 allowed-ips 172.26.0.1/32
2022/03/06 22:45:31 [netclient]
2022/03/06 22:45:31 [netclient] error running command: wg set nm-Tnjg5NP0 peer jKNE8q3T3il/o0QpKLur3wQrM7/kBCTGqdwYFhLym3Y= persistent-keepalive 20 allowed-ips 172.26.0.2/32
2022/03/06 22:45:31 [netclient]
[netmaker] 2022-03-06 22:45:32 sent peer update for node Bernard on network: default
[netmaker] 2022-03-06 22:45:32 ERROR RETRIEVING EXTERNAL PEERS ON SERVER: could not find any records
2022/03/06 22:45:32 [netclient] error running command: wg set nm-default peer S/nyvv6qz6zxUmFdgHcU+K9t8fsoHdsq78Eu2lBD+Fk= persistent-keepalive 20 allowed-ips 10.101.0.1/32
2022/03/06 22:45:32 [netclient]
2022/03/06 22:45:32 [netclient] error running command: wg set nm-default peer 7xtOcbhDg8C0mixP2RRAbZ8fDL38CS6TZrhuhQxK7jU= persistent-keepalive 20 allowed-ips 10.101.0.2/32
2022/03/06 22:45:32 [netclient]

running

wg show

on the netmaker machine, it shows no interfaces...could that be the issue here? :/

the wireguard interface on the server is inside the container docker exec -it netmaker wg show

Ah of course

However

OCI runtime exec failed: exec failed: container_linux.go:380: starting container process caused: exec: "wg": executable file not found in $PATH: unknown

you will have to use the full path to the wg executable

What's the full path? /usr/bin/wg didn't work

not sure right off the top of my head

Hello I am trying to connect a node to my network but I'm getting the error that netclient is already running, when I try the suggested command to stop netclient I get an error cmd not found using Ubuntu 20.04

/sbin/netclient

I just had a look at docker-compose.nocaddy.yml and there's a volume

/usr/bin/wg:/usr/bin/wg

, so I assume that wg should be there? 🤔

Is it possible to create a bash shell to the container to look around?

docker exec -it netmaker /bin/bash

fails as it can't find /bin/bash. Is that expected or is my container very broken? 😅

Aha, I used the docker-compose.nocaddy.yml from the develop branch, where bind m ount to /usr/bin/wg was removed from docker-compose files... is it a bad mix to use that file from the develop branch and using the nm-quick script to install netmaker? I suspect that the wireguard-tools isn't included in the dockerfile.

When uh...is the Dockerfile (that now is the file that adds wireguard-tools) used if I install netmaker with nm-quick.sh? 🤔

Still have this question. Anybody knows something about this?

No this is not currently possible. The next version may allow this but don’t want to confirm this yet.

^ along those lines, is there any way to manually override the wireguard config (or at least the allowed-ips) for the netmaker server node?

I have an app which uses multicast udp packets to communicate between devices. For now I've been using zerotier. Can I achieve the same effect with netmaker?

as far as I know you can't do mutilcast with wireguard.

you can't indeed

zerotier builds a layer 2 network, wireguard is layer 3

So I got a node behind a LTE stick (a rapsi) When it joins the network, ot seems only the netmaker server is able to communicate with it, it's tunnels are working. But, with all other nodes it seems unable to setup the tunnel correctly. Any idea how to even approach debugging this?

I have seen the same issue with an LTE modem

On other nodes it seems like they send packages but hear nothing back