Create build-deploy-ec2.yml

[paulespinosa]

This initial GitHub Workflow, as a manual action, builds and uploads the Home Unite Us front-end application and back-end API web service to the EC2 instance currently used for the Home Unite Us project.

In this commit, the workflow does not cause the uploaded packages to run. That will come after testing and refining this commit.

related #465

[Joshua-Douglas]

Hey @paulespinosa,

Looks great! Have you been deploying manually up to this point? If so, I'm sure this will make your life a lot easier.

I just have a few minor comments and questions.

[Joshua-Douglas]

Is this the successful build log? https://github.com/hackforla/HomeUniteUs/actions/runs/5651696696/job/15310200967

If so, I can't seem to find the tox step. Am I overlooking it? I'm mostly wondering because the api test cases were failing on my machine. I'd expect them to fail here to, unless I'm missing something

[Joshua-Douglas]

🤦‍♂️That link I posted was the PR build link haha.

Have you tried out the workflow and confirmed that it worked? If so, can I take a look at the deployment log?

Thanks!

[paulespinosa]

😊 Actually, I'm currently figuring out how to test it without merging to the main branch.

[Joshua-Douglas]

Have you tried out this workflow yet? It might be worth sending off a build before approving, to ensure the SSH access is working.

Do we have a development EC2 instance? If we did, then we wouldn't have to worry about breaking the current site

[paulespinosa]

@Joshua-Douglas Yes, I've got it in a testable state and have been modifying it. You can see all of the runs in the Actions tab.

At this time, we have just one EC2 instance. It hosts the static landing page for https://homeunite.us/ and acts as the dev/test/demo host for https://dev.homeunite.us. As of this writing, it's hosting an old version of the frontend and the backend is down.

[paulespinosa]

@Joshua-Douglas It looks like the SSH keys aren't configured in GitHub secrets yet. We'll need to check with @tylerthome about that.

[paulespinosa]

Hey @paulespinosa,

Looks great! Have you been deploying manually up to this point? If so, I'm sure this will make your life a lot easier.

I just have a few minor comments and questions.

As far as I can tell, we haven't been deploying code for quite a while. I think Tyler was manually doing the deployments in the past.

[Joshua-Douglas]

As far as I can tell, we haven't been deploying code for quite a while. I think Tyler was manually doing the deployments in the past.

Ah, that is great to know. Thanks. Now I've got a bunch of questions about the version of our deployment and our deployment strategy. I'll ask at today's meeting!

Either way, thanks for getting this started!

[erikguntner]

current working directory "/home/runner/work/HomeUniteUs/HomeUniteUs"

@paulespinosa you may need to set the working directory to app for the end to end tests in the build-deploy-app workflow.

[paulespinosa]

current working directory "/home/runner/work/HomeUniteUs/HomeUniteUs"

@paulespinosa you may need to set the working directory to app for the end to end tests in the build-deploy-app workflow.

@erikguntner Thank you for catching that. I added it back in and it looks like its working again.

[ju1es]

sweet. seeing in the build that it's failing on the service provider controller tests. I think Joshua has a PR that should fix.

I forget if we're working towards ci/cd but either way, can we update the readme with instructions for deployment.

[paulespinosa]

@tylerthome For this workflow, I used the shimataro ssh key action with the following configuration:

      - name: Install SSH key
        uses: shimataro/ssh-key-action@v2
        with:
          key: ${{ secrets.SSH_KEY }}
          known_hosts: ${{ secrets.KNOWN_HOSTS }}

When the workflow runs, it appears that the shimataro ssh key action isn't picking up the SSH_KEY or KNOWN_HOSTS secrets. The error is "Error: Error: Input required and not supplied: key". Am I using the right variable names?

Also, it looks like we'll need to add the .env variables to GitHub secrets because the workflow runs tox as a step. When the tests runs the auth_controller.py module is loaded. This module has a top-level instantiation of a boto3 client that checks for the AWS credentials during initialization. Since the environment doesn't have the AWS credentials, the boto3 client instantiation fails causing the tests to fail.

[paulespinosa]

The workflow is being added to the PR #544 because the changes are related. This PR is being closed.