Added support for a database installer script. (#438)

[jeremyphilemon]

Link the Issue(s) this Pull Request is related to.
Fixes #438

Summarize your change.

• Added a bash script in sandbox/ called setup-database.sh – automates the installation and setting up of the database, including applying migrations.
• The steps for installation were taken from the "Getting Started" Guide
• Uses the quick fix referred to in Latest Postgres docker needs additional environment variables #651 instead of the one in the documentation due to a breaking change in docker-library's postgres image.

It's also helpful to describe why you're making this change.

When installing and setting up the database, there are numerous steps that can go wrong. It would be nice to have an installer script to guide you through the database install and configuration. This same script could also be used to update and apply migrations to existing databases. - @gregdenton

Hence :)

Testing

• From the root directory of the codebase run ./sandbox/setup-database.sh

[linux-foundation-easycla]

The committers are authorized under a signed CLA.

• ✅ Jeremy (0178af0)

[bcipriano]

Only one change to request -- let's add a note/disclaimer to the top of the script explaining that this is experimental and isn't recommended for general use yet.

It's basically a translation of the existing process and that's a great start. As a followup change we'll want to look at integrating this into the rest of the sandbox process and see what changes that entails.

Once the script is proven in that scenario, we can start looking into updating the docs to see how this script can replace the existing manual process.

[jeremyphilemon]

@bcipriano sounds good! I'll make the changes and update the PR!

[jeremyphilemon]

Updated!

So here's my understanding: the sandbox currently uses docker.yml to set up the database. If we're looking to integrate this script into the rest of the sandbox process, which like you mentioned earlier is a translation of the existing process, will we be dropping the usage of docker.yml?

I'm under the impression that sandbox is used to build up a quick configuration using docker and deploy the database, cuebot, and rqd with minimal configuration. Hence the reason to why it doesn't fall in line with the existing process. But isn't the existing process for production?

Would like to have that clarified :)

[bcipriano]

Sorry for the delay here.

That's docker-compose.yml you're talking about, right?

My rough idea was not to drop using that file entirely, but just update the db/flyway sections of it to use this new script instead.

Anyway I don't think it's necessary in this PR -- this is a good start and we can figure out next steps over on the issue.

[gregdenton]

I agree with Brian, lets keep the docker-compose.yml but we can certainly update the file to utilize the script and make the deployment patterns more similar.

[gregdenton]

Thank you for the update!
Changes LGTM.

[jeremyphilemon]

@bcipriano @gregdenton Alright, sounds good! :)