James vms deployment

[JPadovano1483]

@ted-ross Here is what I currently have working for the deployment of the required VMS components:

• Custom helm chart for deploying the management controller - this can be deployed by itself via helm install or with the other components with the helmfile cli
• public postgres helm chart installed via helmfile
• public cert-manager helm chart installed via helmfile
• Custom keycloak chart - this installs the keycloak crds and operator and creates a keycloak instance and optionally a postgres instance alongside it. (The keycloak CRDs and operator are copied into the keycloak-stack helm chart from the official Keycloak docs. This is fine for trying out vms or for development but for production environments, it is probably best to go through OperatorHub).

I am still improving and polishing some things, but wanted to get feedback on the overall current state and setup. One thing I should have asked more about earlier that I mentioned to Andy is that the helmfile cli is a community managed tool, so I am not sure what IBM's policy is on using this. He is going to look into it, but we can pivot away from using that if need be.

The README still needs to be cleaned up and vetted more, but I believe following the "Prerequisites" and "Deploying on a single cluster" sections should get you running. One thing to note is that Keycloak is a bit particular about using a hostname. I found that if I didn't set a hostname and create an ingress, the auth callback would fail. I will keep exploring configuring this for development environments. The steps I took to configure this were:

1. Enable the ingress minikube addon
2. Map a hostname to my minikube IP in /etc/hosts
3. Use this hostname in values/common.yaml -> keycloak.hostname

I would recommend using a fresh cluster or one without cert-manager installed for testing. Feel free to grab me if it would be easier to go through that way.

[JPadovano1483]

Changes:

• removed the keycloak release
• removed automated cross cluster deployment
• updated documentation

I left the original database scripts, yaml, and getting-started documentation for now. Let me know if I should remove these. Currently there are copies of the database scripts in the helmfile directory, but before merging I will make sure there is only one source for these. I think if we want to remove the old yaml, I would say we should keep the scripts in the helmfile directory, but if we want to keep them around I would leave the scripts where they are.

[JPadovano1483]

Fixes #10

[ted-ross]

I think the SQL scripts should be removed from their old location so we don't cause confusion by having two copies.