Page 1 of 1

RevPi Core S and moDIO closing SSH randomly

Posted: 10 Sep 2024, 10:59
by wmatte
Dear community,

Hello everybody! I am still developing things on this amazing platform :) This time I am programming a second RevPi Core S with a moDIO module, and still amazing.

Nevertheless, this time I am encountering a problem whose root cause I cannot understand.

I am programming the RevPi using SSH via a remote connection. So far so good. Btw, randomly, it closes the connection and even if the RevPi is up, is impossible to connect again.

When this happened, I saw that the PWR LED turned red and then green again, making me think it was something connected to the watchdog. I would like to dig a little more into it, but I have no idea where to start.

I take advantage of this problem to kindly ask for a general approach to RevPi debugging, to gain some skills in this direction :)

Thank you again to the community for the support.
Best regards :D

Ps: I am programming it using VSCode!

Re: RevPi Core S and moDIO closing SSH randomly

Posted: 10 Sep 2024, 13:47
by dirk
Hello wmatte and welcome to the Revolution Pi community!
So since the Bullseye OS, the "NetworkManager" service has been active here and controls the network.

So the root cause may be that the IP connection was switched to "fallback". This occurs, i.e. if the DHCP server does not respond or is not available. Then it looks like that:
fallback connection
fallback connection
You can increase the timeout or the autoconnect-retries here:

Code: Select all

/etc/NetworkManager/system-connections/dhcp-eth0.nmconnection 
NetworkManager configuration
NetworkManager configuration
Do you already know the slightly fancier option of making SSH logins without using IP addresses?
https://youtu.be/tQJ4iSPT_LY?si=Ulu14D91AVb8cW4s

Re: RevPi Core S and moDIO closing SSH randomly

Posted: 11 Sep 2024, 08:58
by dirk
Okay but the PWR LED thing here made me think, I hadn't considered this detail yesterday. Hm.
So if the PWR LED turns red and then green again, this is a sign that the communication between the modules is faulty. The I/O modules then go into the safe state and switch the outputs to logic 0. Can you create an SOS report?