I am pretty sure this topic has come up before. But, if I recall correctly, there never was a definitive answer.
Are there instructions for configuring a newly flashed (or prior to saved saved settings) upgraded node from the command line (ssh or telnet)?
For one use case we have 2 dtdlinked nodes. Both nodes are on a rooftop that is very hard to get access to. One of the two nodes was patched V3.0.2 OTA and never came back. The node is now advertising itself as SSID meshNode, but I cannot connect to it. (It is not offering DHCP and manually setting a client to 192.168.1.2 didn't work either). I was thinking I could access it from the linked node and maybe fix the problem - (I would need to change the LAN address, or add a second interface on the working node to do that.).
I got on site today and found the node I thought was wedged in NOCALL mode was actually operational after the upgrade from 3.0.2 (saved settings) to 3.15.1b4. I was able to access it using the dtdlink, reset the password and it came right up. The problem with the SSID meshNode, which had confused me, turns out to be some other unidentified node. This is on the UC Berkeley campus, so who knows what could be floating around.
However, it would still be nice to know how to configure a node "by hand."
wb6tae, There's a perl script, "/usr/local/bin/node-setup" that has been around since linksys days.
usage: node-setup [-a] [-p] <configname>
-a: automatic mode - don't ask any questions
-p: only process port forwarding and dhcp settings
This is certainly in the unknown category since we've not purposely been testing it, although it has had updates this year. I'll defer to others if anyone has used this recently, I have not.
Note, if you have command line access, it means there is a way to restart uhttpd, if it wasn't responding, and then access the UI.
Joe AE6XE
node-setup requires the GUI to have been used at least once in its current state. Some dynamically configured items are only initially set after first save in the web gui.
So no you can do initial setup from CLI at this time in a supported fashion. If it was really stuck (I understand it actually wasn't in this case) it would of probably been in a state where it needed that GUI code to execute to be sure.
On reflection (something probably better done in advance, but, then it wouldn't be reflection)... It is hard to define a case where a new node would need to be brought up without GUI access. The hint to restart uhttpd is a good one, should that ever be necessary.