Follow-up from "Hashed Storage is enabled by default on new installations"

The following discussion from gitlab-ce!29586 should be addressed:

• @brodock started a discussion: (+2 comments)

  some context on specs failure (gitaly log):

  time="2019-06-13T15:54:34+02:00" level=error msg="fatal: Couldn't find remote ref HEAD\\nfatal: the remote end hung up unexpectedly\\n" grpc.meta.auth_version=v2 grpc.meta.client_name=gitlab-web grpc.method=FetchIntoObjectPool grpc.request.deadline="2019-06-13T15:55:29+02:00" grpc.request.fullMethod=/gitaly.ObjectPoolService/FetchIntoObjectPool grpc.service=gitaly.ObjectPoolService grpc.start_time="2019-06-13T15:54:34+02:00" peer.address= span.kind=server system=grpc
  time="2019-06-13T15:54:34+02:00" level=error msg="finished unary call with code Internal" error="rpc error: code = Internal desc = exit status 128" grpc.code=Internal grpc.meta.auth_version=v2 grpc.meta.client_name=gitlab-web grpc.method=FetchIntoObjectPool grpc.request.deadline="2019-06-13T15:55:29+02:00" grpc.request.fullMethod=/gitaly.ObjectPoolService/FetchIntoObjectPool grpc.service=gitaly.ObjectPoolService grpc.start_time="2019-06-13T15:54:34+02:00" grpc.time_ms=35.663 peer.address= span.kind=server system=grpc

(https://gitlab.com/gitlab-org/gitlab-ce/-/jobs/230655211)

rspec ./spec/finders/fork_projects_finder_spec.rb:29 # ForkProjectsFinder#execute without a user 
rspec ./spec/finders/fork_projects_finder_spec.rb:35 # ForkProjectsFinder#execute with a user 
rspec ./spec/finders/fork_projects_finder_spec.rb:41 # ForkProjectsFinder#execute with a member 

before the change to hashed_storage_enabled: true, the failed specs would not try to use repository_pool because this would be false:

  def git_objects_poolable?
    hashed_storage?(:repository) &&
      public? &&
      repository_exists? &&
      Gitlab::CurrentSettings.hashed_storage_enabled &&
      Feature.enabled?(:object_pools, self)
  end

As this now true, we are having a failed spec. I'm still trying to determine if this is an issue or not.

In https://gitlab.com/gitlab-org/gitlab-ce/merge_requests/29586 I've stubbed the feature flag to disable pool repositories. We should investigate if this is a legitimate failure or just a fluke on how we are setting up the tests.

Ideally we should test both with and without pool repositories states.