I have an installation (CerboGX, running v3.66 large) running a script that I’ve put together that optimises off-grid / grid-parallel operation, while also constantly updating grid set-point while on grid-parallel mode. Script has been running fine for almost a year now.
However, since the beginning I’ve been having issues connecting to node-red or its dashboard a few minutes after the system switches to grid-parallel mode. It’s like it’s loading tooooo slowly and node-red or its dashboard eventually time out.
If enough time has passed since the switch (maybe, half an hour, I can’t tell exactly), or when system operation switches to off-grid mode (yes, the node-red script still works to make the switch), node-red connectivity eventually resumes.
While the node-red-related connectivity is lost, everything else continues to work fine. VRM shows as online, remote console works fine, I can access the remote console locally by going to http://venus.local, but I can NOT access the node-red interface remotely or by accessing http://venus.local:1881.
I get no error. Just browser timeout eventually. Accessing http://venus.local IS indeed local. Of course I tried using the local IP address of the Cerbo, as mentioned Remote Console works fine, node-red processes fail to launch.
Thinking of logging into the Cerbo using SSH, in order to keep an eye on wether ther is extra CPU/memory usage when the issue occurs, if top or a similar command is available on the Venus.
Anyone knows of a way to remote connect to the Cerbo via SSH? I tried enabling “superuser” and “remote support”, which informs me of a tunnel IP, but I still wasn’t able to log in, hence I guess it’s there solely for Victron engineers.
To allow me to access my kit remotely I have installed Tailscale on my routers, both at home and on the boat, as well as on my PC. laptop, and tablet. This allows me to access from anywhere. You can run Tailscale on a lot of devices, so it may be worth looking at putting a gateway on your network. Tailscale is free for non-commerial use.
Turned out my node-red installation running slow or completely not responding on various instances was due to a combination of a few scripts using too much context data and a 15-min Chart that I had set up, that would use up to 3600 points and I guess it was too much for the Cerbo.
Anyway, completely reworked the code and disabled the Chart and now I no longer have delays or inability to connect.