Provider cleanup

[stevenpost]

This PR cleans up the provider.

A detailed list of changes:

• Remove deprecated SSL options (use TLS instead)
• Support pure yaml config in provider
• Refactor the replicaset provider to improve maintainability

[stevenpost]

I decided to split this into 2 parts. A 'safe' part, where it is just cleanup (#726), but no real changes in behaviour and another 'unsafe' part with changes in behaviour. The second may provoke a bit more discussion and is this PR.

This PR will be split by @stevenpost

[stevenpost]

I did some refactoring of the replicaset provider to improve maintainability.

My current test setup and requirements:

• a 3 node cluster
• PuppetDB
• Puppet needs to run only once to fully set up a node (and extend the replicaset)

Test procedure:

• Spin up 3 empty VMs
• Run Puppet on all three nodes sequentially

The node will only have itself in the replicaset.
The second and third nodes will extend the existing replicaset.

[h-haaks]

Is it possible for you to add your setup in a beaker hosts file under spec/acceptance/nodesets?

[stevenpost]

Is it possible for you to add your setup in a beaker hosts file under spec/acceptance/nodesets?

Not sure how to go about that, but I'll give it a shot.

[h-haaks]

Hmm, this a994a6c was not what I had in mind :)
I was talking about creating a beaker hosts yaml file in the spec/acceptanse/nodesets folder with 3 host.
This file could then be used with bundle exec rake beaker

I suppose its better to do this in a separate PR :)

[stevenpost]

Hmm, this a994a6c was not what I had in mind :) I was talking about creating a beaker hosts yaml file in the spec/acceptanse/nodesets folder with 3 host. This file could then be used with bundle exec rake beaker

I suppose its better to do this in a separate PR :)

I moved that commit into a separate PR (#729). I think I understand what you mean, no clue on how to do that, though. Also, I don't see what will be actually tested. Can you expand on that (perhaps in the new PR)?