Now Magnum is using podman and systemd to manage the k8s components.
In cases where the nodes pull images from docker.io or another
mirror registry with high latency, some of the components may take long
time to start, which is causing timeout when bootstraping k8s
cluster for fedora atomic/coreos drivers. This patch fixes it by
adding TimeoutStartSec for the systemd services.
Task: 37251
Story: 2006459
Change-Id: I709bac620e4ceec1858672076eb0aef997704b62