Prerequisites for running ansible inside the appliance

[carbonin]

This PR creates the postgres database that will be used for ansible on the appliance when we create our database.

Additionally this PR adds the host and password to database.yml even when we are running a database locally. This was done so that the password is accessible at configuration time to provide to the ansible process so that it can contact the database.

https://www.pivotaltracker.com/story/show/137048481

[miq-bot]

Checked commits carbonin/manageiq-gems-pending@6a90c9c~...16baf81 with ruby 2.2.6, rubocop 0.46.0, and haml-lint 0.19.0
3 files checked, 0 offenses detected
Everything looks good. 🍪

[carbonin]

An alternative approach here could be to execute this stuff when we do the configuration and generate and save the db password like we do for the other keys and passwords (ManageIQ/manageiq#13584). This way we have a separate place to deal with any password rules or special character restrictions that the ansible side may have.

Thoughts @gtanzillo @jrafanie ?

[jrafanie]

An alternative approach here could be to execute this stuff when we do the configuration and generate and save the db password like we do for the other keys and passwords (ManageIQ/manageiq#13584). This way we have a separate place to deal with any password rules or special character restrictions that the ansible side may have.

Thoughts @gtanzillo @jrafanie ?

I don't know, maybe I'm missing something but it feels like doing this as configure database time is exactly when it needs to happen, when creating the internal database. I'm open to other ideas but I don't see why we'd do it differently.

[jrafanie]

Other than adding the host and password that was previously not in the yaml, it seems fine. And @kbrock signed off on the host/password, so I'm good with this.

[gtanzillo]

I'm in agreement with @jrafanie. Unless there's some reason (other being uncomfortable with adding host and password to the database.yaml file) why we shouldn't or can't create the DB for ansible, I think it should be done at the same time as the vmdb DB. So I'm good with this. anyway, we already save the password when the DB is not local.

[kbrock]

I liked having nil for the hostname, which uses a local socket.

but the main benefits seem to be:

• gets us closer to having a common database.yml for all appliances in a region.
• makes failover easier when teetering between localhost and remote host (which needs a password).
• stores db password for use by Ansible.

[gtanzillo]

👍