fix restart_services method in pool.py

The "restart_services" method in the pool plugin is causing us more harm than good. It's restarting collectd/rrdcached more than once (this is painful on large systems)

It's also abstracting away very simple CLI commands that are being run that isn't necessary.

There is also another improvement that can be made (while I'm here) to be more verbose and log more informational messages to the end-user when a pool is being created/imported.