[CentOS] LXC containers - systemd takes a long time to start

I've got a CentOS7 server (fully patched) and in it I'm running several 
LXC containers.  I've noticed that whenever I try to launch a service 
via `systemctl start <service_name>` soon after LXC boot, it will take 
~5 minutes before the prompt returns.  After that initial delay, 
starting/stopping/restarting services happens immediately.  It might be 
that this happens because the server is running a large number of (>20) 
containers.  On next reboot, I'll see if it has the same problem when I 
have just one container.

In the meantime, anyone else seen similar behavior?