akash-provider: move run.sh into a standalone file

[boz]

So that it can be linted and tested outside of a pod.

https://github.com/ovrclk/helm-charts/blob/b52f46016de4a4516c6a4711fecead0cb58d7bbf/charts/akash-provider/templates/configmap-boot.yaml#L7-L69

[arno01]

Here is a PR #41 that moves run.sh into a standalone file.

It is till in the form of template though.
Not sure if we want to go beyond this point further and make it entirely non-template-based as one won't be able to test it straight away, unless setting all the environment variables correctly which are also templated here https://github.com/ovrclk/helm-charts/blob/c10fa285da5c72a396e28ad61615632b029048d5/charts/akash-provider/templates/deployment.yaml#L60-L154

run.sh is currently having two parts where it needs to be templated:

• provider.yaml file generation (we are passing these via https://github.com/ovrclk/helm-charts/blob/c10fa285da5c72a396e28ad61615632b029048d5/charts/akash-provider/values.yaml#L61-L77 or directly on the CLI via --set attributes[N].key/value= arguments) => akash-provider: generate attributes in a config map. #14
• /bin/akash tx cert generate server provider.{{ .Values.domain }}

We could probably use another two template-based configmaps and let run.sh script read those files instead, but, I think that will be too over-engineered.

Upd:

And, ideally, the docker image should only run akash provider run (get configs via the env. variables only), and configmaps only to deliver the config files (provider.yaml, etc).
But since this run.sh script does more things such as provider attributes update, certificate registration, etc.. these steps should probably really be separated as suggested here #9 (comment)