Allow managing logrotate defaults

Just for cosmetics

Following the template-formula example.

add Scientific Linux support

No need to reload the service on configuration change

Updates logrotate.conf for Debian

/var/log group in Ubuntu is syslog. This formula breaks
Ubuntu logrotate.

path (if mapping) need to be list now

Support multiple log path

Added a muktipath example on pillar.example file

Added a multipath example on pillar.example file

Fix job defaults

The iteritems dictionary method has been removed in Python 3 and using
items() is typically no problem in our case due to the fact that
dictionaries are rather small.

This fixes #15

This allows users to target this states to minions for which no pillar
has been defined (yet).

add logrotate.conf for suse

fixes issue #5

splits config file name and log file path, so you configure both.
This fixes #5.

Change the default service name to 'crond' for RHEL based systems.

Moving logrotate directory setup from jobs.sls and config.sls to init.sls

Remove quotes

None of these quotes are actually needed.

Added opensuse support

added Fedora to map.jinja

it's crond service on fedora that initiates logrotate

Update readme and pillar.example

add support jobs