Changed it to draft because I want to debug the failed test in Admission Webhook Tests run.

I could not increase the resources inside the VM (what in the runner is the max I could for nb install and the create operation to succeed).
As I wrote:

One of the admission tests is still failing (Delete operations - Should Allow - bs delete), I'm not able to reproduce it on my minikube, had to push to GH to see the workflow result. The fix I'm suggesting seems to reveal an internal race condition in the core. When the tests failed the bs-name backing store mode is DEFAULT_RESOURCE and it cannot be deleted (the cases where it passes as I saw was with MANY_STORAGE_ISSUES mode).

Today the default resource is set in the core in the flow of (inside pool_server): _init -> update_account_default_resource -> get_optimal_non_mongo_pool_id (when there is a moment where the bs-name backingstore is OPTIMAL this resource is set as default resource (of the account admin@noobaa.io).

What we decided to do was to use an RPC call that allows us to change the default resource of the account explicitly before trying to delete it.

@dannyzaken I run the admission tests from the PR and it passed (more than once), you can see it here.

Only the first commit of this draft will stay, and the other changes are in PR #1046.