Auto-generate a stateful erlangCookie

[colearendt]

What this PR does / why we need it:

Discussed in #78

Which issue this PR fixes

(optional, in fixes #<issue number>(, fixes #<issue_number>, ...) format, will close that issue when PR gets merged)

• fixes Changed auto-generated erlang cookie causes cluster restart issues #78
• fixes Fix broken CI on main #88

Special notes for your reviewer:

• Need to make sure that NOTES.txt and the new "stateful" secret generation does not error on various types of inputs (I intend to do more testing here)
• Some design choices could be evaluated. For instance:
  • Simplify the input by passing the global context (so Namespace can be determined) and hard-coding the secret name
  • Whether or not to use NOTES.txt the way we have, or whether a generic message would be more helpful (i.e. is it too verbose? Should it be in a different location?)
  • Whether NEWS.md seems like a helpful convention
  • Whether we should be using index to protect against missing keys like I did in allow uuid being created and managed by kubernetes #73
  • Whether this is a desirable approach or erlangCookie should be refactored more heavily

Checklist

[Place an '[x]' (no spaces) in all applicable fields. Please remove unrelated fields.

• Chart Version bumped
• e2e tests pass
• Variables are documented in the README.md
• Chart tgz added to /docs and index updated

[colearendt]

Interesting CI error... Kernel pid terminated (application_controller) ({application_start_failure,kernel,{{shutdown,{failed_to_start_child,net_sup,{shutdown,{failed_to_start_child,auth,{"Bad characters in cookie",[{auth,ini

I think I may have messed up quoting...

EDIT: fixed!

[colearendt]

For the record, I haven’t had a chance to test this locally, but I am pretty sure it will resolve #88 because the “helm upgrade” cannot proceed with a different erlang cookie

[colearendt]

Alright! Finally got around to testing this! This does indeed fix the upgrading issue that is breaking CI (#78). It also works for this upgrade because it detects the previously autogenerated value and persists that! (I had an issue with base64 operations that was breaking the PR CI).

FYI @willholley

[colearendt]

@willholley is this good to merge? (I don't have access to do so 😄 )