How to handle failures

dpfrey wrote on Thursday, March 02, 2017:

What is the best way to deal with non-recoverable failures in FreeRTOS? Is the solution to just put configAssert() calls into the code? I intend for these checks to remain in production and not just to be for debugging during development.

rtel wrote on Friday, March 03, 2017:

Would need more information, but ‘non recoverable failure’ would seem to
indicate that the failure cannot be recovered from, so the best thing to
do is very dependent on your application.

If it is safe to reboot, then reboot.
If it is not safe to reboot then take actions that place the system into
a known state.

You can use the FreeRTOS+Trace tool to keep a rolling log of kernel
activity. If you stop the log when the error occurs it might help with
a postmortem.