Rnsh over transport RNodes

[dannydelcaribe]

Greetings!

Recently I have been doing some range testing with RNodes and trying to see how far an rnsh connection I can get by using transport nodes to bridge two end devices.

My setup consists of 3 raspberry pis. Each with a Heltec v4 RNode. I have configured two of them to be endpoints. While on of them have transport enabled = yes. The goal is to extend the range of the rnsh connection.

Diagram:

RNode_A <====> RNode_B <====> RNode_C
.....full...................transport.................full.......
.initiator.................gateway...........announce 10s

I can rnsh between RNode_A and RNode_C by going point to point no problem. The problem comes when I get A and C too far apart and attempt to close the gap with B. In this configuration I get the "Establishing Link..." message on the initiator, but the link never happens.

I mention the "Establishing Link..." message because it is consistently an indicator I can see the serving end. At least when going point to point.

My current estimation is that perhaps a transport RNode cannot handle an rnsh link and this is therefore impossible to do. I would really like this to work however and if maybe can shed some insight, it would be much appreciated.

[acehoss]

Hi @dannydelcaribe,

I haven't used rnsh much over LoRa--I have, just not much--but I have used it a fair amount over multi-hop links and I'm quite sure others are using it for multi-hop LoRa links.

I don't know what your technical skill level is or how familiar you are with Reticulum, but I'd recommend testing with one of the Reticulum-provided utilities like rnx to see if you have similar problems with that approach.

What other information can you share about your setup? Are the nodes all located in the same room for testing? What LoRa settings have you configured? Are you able to capture any logs on any of the nodes while attempting to establish a session?

[dannydelcaribe]

Thanks @acehoss, I haven't been able to test with rnx yet. I've been a little busy, but I will soon!

To give you an idea of my physical setup, and I've done a few different ones. I set up a LoRa node connected to a Raspberry Pi hosting (serving) rnsh. I then took my laptop and connected to the pi via rnsh. On the pi I simply ran rnsh -l -n and on my laptop I ran rnsh <listener of pi>. This worked normally and both radios had the same parameters:

[[RNode1]]
    type = RNodeInterface
    enabled = yes
    frequency = 915000000
    bandwidth = 500000
    spreadfactor = 5
    codingrate = 5

This configuration gave me a total bitrate of 62kbps.

After successfully connecting at short range, I moved a node away from the other until I could no longer send the ls command. Additionally, I enabled line-interactive mode ~L.

Once I couldn't interact with the pi, I placed a transport enabled, gateway mode node in between them. The parameters on this transport node were the same as above, with the exception of mode = gateway line. This did not seem to work.

I should note that my rnsh server (raspberry pi) was set to announce only once on startup. I'm not sure if the transport/gateway node needs to hear an announce from the server to work properly. However in this case it did not hear one. It would have only heard my initiator request.

In the future I will test this with rnx and also capture some logs.

[acehoss]

My Reticulum knowledge is fairly faded at this point, but I think you need to make sure you're advertising somewhat regularly. I don't recall how path discovery works, but when I run rnsh listeners I do set them up to send out periodic advertisements.