* Move the config logic into a defined type
* This will allow us to have multiple instances
of redis on one machine
* Example code:
```
# Install redis package
class { '::redis':
  default_install => false,
}

redis::instance {'redis1':
  port     => '7777',
}

redis::instance {'redis2':
  port     => '8888',
}
```

* New images doesn't have the locale issues

* Fixes endpoints to be https
* Fixed a long time ago (https://github.com/stahnma/puppet-module-epel/issues/34)
* Updates spec_helper_acceptance and metadata

* Adding deprecation warning
* First steps towards the work in #152

* Allows automatic generation of docs
* Adds rake task for doc generation

* The original behaviour seemed a little brittle to me, so I asked for some help
* This new refactor seems a lot cleaner, and deals with the handling of getting a nil response a lot cleaner
* Specs edited to add in two types of disconnection errors: getaddrinfo for non-existent DNS and ECONNREFUSED for when connection refused
* Credit: https://codereview.stackexchange.com/questions/163727/redis-lookup-with-a-default-value/

* Add blacksmith tooling for tagging

* Specifies a default value if the value is not
found or theres a failure to connect to redis
* Adds spec and acceptance tests for new default
value

* redisget() doc update
* getvar_emptystring() doc update

* The existence of /var/run/redis causes the startup of redis-server via
systemd to fail on debian systems with redis-server >= 3.2.8-2:

```
May 18 19:17:54 debian-8 systemd[1]: Starting Advanced key-value store...
       May 18 19:17:54 debian-8 systemd[803]: Failed at step RUNTIME_DIRECTORY spawning /bin/run-parts: File exists
       May 18 19:17:54 debian-8 systemd[804]: Failed at step RUNTIME_DIRECTORY spawning /usr/bin/redis-server: File exists
       May 18 19:17:54 debian-8 systemd[1]: redis-server.service: control process exited, code=exited status=233
       May 18 19:17:54 debian-8 systemd[1]: Failed to start Advanced key-value store.
       May 18 19:17:54 debian-8 systemd[1]: Unit redis-server.service entered failed state.
       May 18 19:17:54 debian-8 systemd[1]: redis-server.service holdoff time over, scheduling restart.
```
* It will actually eventually start, but the first run will fail
* The default unit file for redis-server in debian jessie-backports
contains `RuntimeDirectory=redis`, see [this commit](https://github.com/lamby/pkg-redis/commit/1cecea5abcb2ce05beacd4347977634d06c59cef).
* Adds logic to not manage the /var/run/redis directory if systemd present

When creating new files/directory/symlink, puppet asumes that it's a file
to get the default context. This is wrong as the default context depends
on the file type. See https://tickets.puppetlabs.com/browse/PUP-7559

For the redis case, when creating /etc/systemd/system/redis.service.d
it's setting the context redis_unit_file_t, which is wrong. It should
be systemd_unit_file_t. This has two efects, it's setting the directory
to a wrong context on first run and it's breaking the idempotency of
puppet-redis.

With this patch, puppet-redis is ignoring selinux context for the directory.
Note that the default context is applied automatically by the OS, so it
will be set to the right one.

* With the previous approach, error on 1.8.7 Ruby:
```
Error: Failed to parse template redis/redis.conf.2.4.10.erb:
  Filepath: /etc/puppet/modules/redis/templates/redis.conf.2.4.10.erb
  Line: 87
  Detail: undefined method `<=>' for nil:NilClass
 at /etc/puppet/modules/redis/manifests/config.pp:91 on node master
```
* For ruby 1.8 compatibility
* Should fix older Puppet on systems that have 
1.8.7 Ruby (RHEL6/CentOS6)
* We're not officially supporting 1.8.7 as it's 
EOL, but since this is a minor change, isn't too 
much of a maintenance cost, lets do it! 💃

* If the redis service is unmanaged, like so:
```
redis::managed_by_cluster_manager": true
redis::service_manage": false
redis::notify_service": false
```
then you get a compilation failure as it tries to
look for the redis service to notify:
```
But now an 'include ::redis' gives me:
Error: Could not find dependent Service[redis] for Augeas[Systemd redis ulimit] at /etc/puppet/modules/redis/manifests/ulimit.pp:44
```
* In fact, we don't need any service notification 
in the redis::ulimit class, as it's already in the
redis::config class, which notifies the service 
class anyways
* Adds spec for this use-case to avoid regression

Closes #197

* Changing Travis back to Trusty

* Improves ulimit configuration

* Moves to a dedicated class to DRY up config.pp
* Adds support for systemd systems:
 - /etc/systemd/system/redis.service.d/limit.conf
* Add support for redis when used with a cluster
manager such as pacemaker or rgmanager
* New parameter `managed_by_cluster_manager`
* This manages the file under:
```
/etc/security/limits.d/redis.conf
```
* Closes #130

* This is done at a class level in init.pp
* Do not need to specify individual service

Release candidate for the 3.0.0 release

This will be the last release that supports 
Puppet 3.X

Version 4.x of the module onward will remove 3.X support

* Remove point releases

* We say we support 1404, but aren't testing
* But 1404 has an older package, and certain 
config settings aren't present (See #111)
* Add 2.8.X config file and logic to fix that
* Add to CI, add docker and vagrant SUTs

* Allows modification of common redis kernel
tweaks and requirements
* The ones that antirez (redis creator) mentions
and that are mentioned specifically in redis 
when running it in the logs are:
  * Disable THP: https://redis.io/topics/latency
  * Enable Overcommit Memory: https://redis.io/topics/admin
  * somaxconnL http://download.redis.io/redis-stable/redis.conf

* redis-sentinel is present in upstream repos
* It's also present for 16.04/xenial onwards
* So lets only manage that package if we're on
Ubuntu newer or if we're managing the repo

* Location on Debian-like systems is `/etc/redis/sentinel.conf`

Instead of anding all the commands together, we can pass them each to
bundle exec, which achieves the same results.

* Ignore more of the usual bits

* (style) Remove unused exceptions for puppet-lint

* (style) puppet-lint to ignore lines with more than 140 characters

* (style) README to comply with style

* Redis 2.4.10 is old in software terms (2014)
* Redis will refuse to start 
* The amount of logic that would have to be added
to the template to isolate new config values would
be hard to maintain
* For now, lets just have a CentOS 6 specific erb
* Adds spec to check also
* Changes acceptance test for sentinel to not
run on CentOS 6
* Adds function for Puppet 3/4 compatibility with the fact

Add the redisget() function which can get a value for a specified key
from a specified redis URL.