this code isn't super complicated, but it's the current #4 crasher for service-failures (and we fixed #1 & #2), and the current logs aren't sufficient to figure out why it's failing.  we could add `logger` calls to try and help track it down, but really this needs a better language so we can have sane error checking & logging.

example http://crash/58038c35eb612bcc:
2018-08-17T09:38:00.531386-04:00 INFO periodic_scheduler[13802]: cros-machine-id-regen: running /usr/sbin/cros-machine-id-regen -r periodic -t 21600
2018-08-17T09:38:00.672358-04:00 WARNING kernel: [14913.133490] init: cros-machine-id-regen-network main process (13775) terminated with status 1

all other possible errors were thrown away :/