Added extra configuration flags for meta/clustering update. Changed default storage engine to tsm in preparation for the 0.10.0 release.

Modified clustering configuration slightly so that service is fully stopped when purging peer metadata.

Assure the influxdb_configuration_dir path exists before trying to write configuration file to it

Hard-coded hostname and port replaced with Ansible variables.

Uncommenting storage engine from configuration, setting the default to the b1 engine until tsm is ready for prime-time.

Removed extraneous debug references. Clustering is now fully-supported with 3+ nodes. Slight modification to default directory permissions, and default influxdb configuration file.

Added option installation URL variable to vars, where InfluxDB will be installed via URL instead of through the package repo.

Updated clustering logic to add 3+ nodes. Added a few more checks to ensure clustering is successful (and fail if not).

Added is_vagrant variable, modified defaults file to modify logic depending on whether running on vagrant.

Cluster Control Update

Added extra flag to control whether InfluxDB should be started on install. Templates use 'influxdb' group when joining cluster, so all influxdb instances that need to be clustered must be part of that group (previously it was using all hosts).

Small refactoring updates and bug-fixes.

Updated Vagrantfile to use public network (required for clustering). Will break on non-Mac machine though.