More changes to make logrotate hourly

Changes to tracking.log setup to match production

Changes referenced from @Feanil's suggestions

Typo

explicitly create /etc/logrotate.d/hourly directory

remove logrotate directory in templates

Fixing bugs, Ansible doesn't like multiple module calls in one area

Using special_time=hourly instead of hour=*

Adding back subdirectory layout for templates

Manually dropping file for cron.hourly

Typo

Changes to template, creating logrotate script to run hourly

Bug fix

Making sure old tracking.log config is removed

Adding rds control libraries

Jarv/add sudoers for jenkins

adding tag for env and auth json creation

add default

Addition of production tcp socket configuration

Fix flag name typo.

adding bastion role and task to setup access to the dbs from it

Jarv/provision bare ubuntu

Parameterize machine names in edx-west playbooks

Have elastic search start, but not restart each time

Added CAS settings to edxapp

Unicorn branch merged, reverting forum version to master

Added a validation on the machine variable, expected to be passed in
from the command line now, so that the worker play only works on
machines named "util" and the app play on machines named "app".
Otherwise it would be too easy to mix those up.

Now that we're using -e command line options, it's easier to use this to
specify the "modify_db" variable than prompt_vars.  Having every install
block on this question before was such a hassle!  Now we just won't
migrate unless you explicitly say so on the ansible command with
"migrate_db=yes" or "migrate_only=yes".

Conflicts:
	playbooks/edx-west/carnegie-prod-app.yml
	playbooks/edx-west/carnegie-prod-worker.yml
	playbooks/edx-west/cme-prod-app.yml
	playbooks/edx-west/cme-prod-worker.yml
	playbooks/edx-west/stage-app.yml

Conflicts:
	playbooks/edx-west/prod-app.yml
	playbooks/edx-west/prod-worker.yml
	playbooks/edx-west/stage-app.yml
	playbooks/edx-west/stage-worker.yml

Spell things correctly.

Feanil/always update cache prefixes