I've been playing with the very-sleepy-demo for a few days now.
I am noticing that after some time in 'very sleepy' mode, my units are resetting.
I have 2 Sensortags and a 6lbr border router on a linux machine.
The linux machine is running a script to observe the coap resources (using python txthings/twisted.internet).
Since I am investigating when things get reset, the very_sleepy_mode configuration happens manually before each test.
What typically happens is one of 2 situations:
a) the coap script will report both sets of sensor data every 30 seconds
b) one unit resets into 'normal' mode, then the coap script will report just one set of sensor data,
c) The other unit is normal/idle until I try to put it back to 'very sleepy'.
2 - sometimes the coap script loses its own marbles, and I have to kill and re-run the script manually. The result is that both tag observations do come back.
Replicating either of these cases is very easy. 100% repeatable in my environment.
I am thinking of running this with the uart logging to see if maybe the watchdog is firing.
Does anyone else have any debugging suggestions?
TI forums have been promising (for months) a tutorial on how to set up Contiki + IAR.
I have not seen this tutorial get posted yet.
Perhaps Señor Dunkels and the thingsquare guys can share how they did it? : )
I have created a port of the very sleepy example to ship ALL sensor data + RSSI packed in an octet-stream, however I have abandoned that in favor of the 'out of the box' demo provided with contiki. It is far easier to ask for help with the demo code.