Using curl/tar.gz files for ocp4-helpernode and ocp4-playbooks instead of git

[torwen1]

I have opened this topic related to #89.

I had a customer with exactly this request, because they could not have a own GitHub server to host clones from ocp4-helpernode and ocp4-playbooks.

I solved it by introducing a tag in var.tfvars (git:). I also changed both helpernode.tf and install.tf the way, that I moved the original git clone in 2 new null_resource definitions running before the. One was checking, if the git: tag was at the beginning of the URL and used the original git clone code.
The 2nd one checked, if it is starting with http: or https: and used curl to pull the packages and uncompressed it. I left the git checkout, to support the helpernode_tag and playbook_tag.

The customer just needed to prepare the tar.gz/tgz by using git clone on one of the clients with internet access, simply packing the whole directory in a tar.gz and uploading it to the HTTP server.

An alternative to the git: tag was to introduce a new variable in var.tfvars to switch between git and curl. But we decided against it, because it could confuse users to think, that all packages (like helm, etc) are downloaded from a GitHub server.

I have the code here, just need to fine tune it with the results of this discussion.

[torwen1]

Closed because it is no longer applicable.