Rosserial_Python Node Unresponsive
In my robot I have two Teensy 3.5 microcontrollers set up as nodes using rosserial_python. Each controls a number of actuators and reports on their positions. One of these nodes will occasionally and randomly completely stop responding with no indication as to why. If you publish a command to its topics no errors will appear but the node will do nothing with the command. If you echo one of its publishers it will just sit there blankly until you kill it, even for topics that publish several times a second. The logs show nothing weird going on, but to get it to do anything you have to kill the process with CTRL+C, then after a while it will say "escalating to sigterm" and the node will finally exit. If you relaunch the node it will resume working as usual.
I'm using ROS Kinetic on Ubuntu 16.04. These issues have occurred across multiple computers and multiple microcontrollers so it's not a hardware issue. I never have this problem with the other node despite it publishing much larger messages such as odom.
I cannot cause the issue to happen. I've killed and relaunched the node, cycled power, left it running a long time but nothing seems to impact how often it happens It is very intermittent seems to happen every couple days but when the robot is supposed to functional without interaction on startup it's very frustrating.
If anyone has any idea as to why this may occur I'd be grateful.
