simulink sensitivity analysis stops using multiple cpus, not running parallel jobs correctly

i'm running a simulink model with the sensitiity analysis tools in R2019b via gui. running ubuntu 18.04 on a 24cpu workstation.

in the options, there's a parallel tab. i have parallel options turned on. my local clusters are verified, etc.

when i initiate the simulation, the parallel pool is populated and the simulations run in parallel for ~1--1.5 hours.

if the pool timeout is set to something reasonable like 1 hour, the pool will close during the simulation.

if it's set to something long like 6000hours, the pool does not close during this time. instead, the workers idle ( each sitting at ~1--4% cpu with the correct memory footprint for the process)

same behaviour with 4, 6, 12 workers and presumably others in between which i did not test.

attempts to interrupt/halt the evaluation by clicking on the gui button does interrupts the operation, but only after at least dozens of other simulations are performed which can take several hours.

no errors are returned in the console/terminal/command window.

i have zero idea how to diagnose the problem. any insight or suggestions are appreciated. thanks,