Ok, so Ive been stumped by this problem for the last 3 days, and my exploration is reaching its limitations...
The key idea is that at any moment in time, the automation controller could theoretically have any number of threads open to any number of node machines running any number of applications in any number of parallel/serial combinations. Upon client error, it is EASY to stop any further commands from being sent, but this still theoretically leaves all the current commands untouched. Short from sending a reboot command to every node, is there some way I can force-stop the execution of anything that was effectively started by the controller (children, grand-children, etc)?
Any ideas?