This morning our node-red dashboard became inaccessible via VRM. “Could not connect to the GX to download the requested resource. Check your network settings or switch to a different connection and try again.”.
The remote console works fine so there is no actual network issue. Local access to both the console and node-red also works fine.
The device lists says the gateway was seen ‘a minute ago’. Node-red access is set to ‘full’ in the console settings.
The gateway is a Cerbo GX MK2. Other gateway status info says:
| VRM portal mode | Full |
|---|---|
| Remote Console - New UI | Offline |
| Remote Console - Classic UI | Enabled but not online Learn more |
| Remote support | Disabled |
The GX was last rebooted 257 days ago.
Not sure why it says ‘offline’ for new UI above. If I click on ‘Remote Console’ it appears as expected. But so far as I can tell all that status is normal and the GX is connected to VRM as expected. The advanced dashboard is showing all the usual logging.
In the local console under “VRM Online Portal” it says ‘Connections status: No Error’
In trying to work out what’s up I realise that I don’t actually know how the GX sets up the tunnel(s) to enable connections back in from anywhere on the net. It’s been so pointy-and-clicky that I’ve not had to look. Is there one tunnel or several (one for the console, one for node-red, one for signal-k?)
sshing into the GX I see the process:
{setup-tunnel.sh} /bin/bash /opt/victronenergy/ssh-tunnel/setup-tunnel.sh
Is that script responsible for setting up the node-red tunnel, or only the remote console tunnel. I guess they don’t share a tunnel otherwise they would either both work, or neither work.
The info at the top of the script suggests it’s actually for setting up a remote support tunnel.
How does one debug this? It’s tricky without a mental model of how it is supposed to work. I’m fairly sure that just resetting the GX would make things work, but I’d like to understand what’s going on.